JP4785717B2 - Infrared imaging device - Google Patents

Description

  The present invention relates to an infrared imaging device using an infrared solid-state imaging device.

  As a thermal infrared solid-state imaging device as an example of an infrared solid-state imaging device, one in which pixels are two-dimensionally arranged, connected by a drive line for each row, and connected by a signal line for each column is known. In this image sensor, each drive line is selected in turn by the vertical scanning circuit and the switch, and the pixel is energized from the power source through the selected drive line. The output of the pixel is transmitted to the integrating circuit via the signal line, integrated and amplified by the integrating circuit, and sequentially output to the output terminal by the horizontal scanning circuit and the switch (Patent Document 1, Non-Patent Document 1).

Japanese Patent Laying-Open No. 2005-214639 (paragraph 0003) Ishikawa et al., “Low cost 320 × 240 uncooled IRFPA using conventional silicon IC process”, Part of the SPIE Conference on infrastructure Technology and Applications XXV, published in April 1999, Vol. 3698, p. Pages 556 to 564

  Even a slight variation in the peak value of the vertical drive pulse applied to the drive line of the infrared solid-state image sensor is superimposed on the pixel output, and further amplified by the integration circuit, so that the brightness of each horizontal scanning line varies randomly on the screen. It is visually recognized as horizontal noise.

This invention
Imaging means having sensitivity in a predetermined wavelength range;
In an infrared imaging device including an imaging unit that forms an image of a predetermined wavelength region component on an imaging surface of the imaging unit,
Digital conversion means for converting an imaging signal output from the imaging means into a digital signal and outputting a digital imaging signal;
Line average value calculating means for calculating an average of horizontal effective pixel portions for each horizontal scanning period of the digital imaging signal output from the digital conversion means and outputting a line average value;
For each line, a line reference value calculation means for calculating a frame direction average of the line average value and outputting as a line reference value;
Offset calculating means for calculating an offset value for each horizontal scanning period from a level difference obtained by subtracting the line average value output from the line average value calculating means from the line reference value output from the line reference value calculating means; ,
Delay means for delaying the digital imaging signal output from the digital conversion means by a time for processing in the average value calculation means and the offset calculation means;
There is provided an infrared imaging device comprising: an offset correction unit that performs level correction by subtracting the offset value calculated by the offset calculation unit from the digital imaging signal delayed by the delay unit.

  According to the present invention, by calculating the level fluctuation for each horizontal scanning line and correcting the signal level of the imaging signal, the signal level for each horizontal scanning line becomes constant and stable on the screen. It is possible to suppress horizontally-drawn noise in which the luminance of the light fluctuates randomly.

Embodiments of the present invention will be described below with reference to the drawings.
Embodiment 1 FIG.
FIG. 1 is a block diagram showing the configuration of the infrared imaging apparatus according to Embodiment 1 of the present invention.
The illustrated infrared imaging device includes a lens 1, a shutter 2, an imaging device 3, a subtraction circuit 4, an amplification circuit 5, an A / D converter 6, an attenuation circuit 10, an addition circuit 11, and an update circuit. 12, a frame memory 13, a feedback D / A converter 14, a timing generation circuit 17, a shutter control circuit 18, a delay circuit 7, an offset correction circuit 8, and an output D / A converter 19 , An output terminal 9, an average value calculation circuit 15, an offset calculation circuit 16, and a reference value generation circuit 23.

  First, an operation in a state where the shutter 2 is opened as a standard operation state will be described. The infrared rays emitted from the subject are collected by the lens 1 and imaged on the image sensor 3 through the shutter 2 in the open state. In the image pickup element 3, elements for detecting infrared rays are arranged on a two-dimensional plane, and a signal that changes in accordance with the infrared intensity is obtained from each element. The infrared detection elements arranged in the imaging element 3 have sensitivity in a wavelength band of about 8 to 14 micrometers.

  The subtraction circuit 4 has only a signal component obtained by subtracting the fixed pattern noise AFPN obtained by analog conversion of the digital fixed pattern noise (FPN) stored in the frame memory 13 from the imaging signal SN that is an output signal of the imaging device 3. The signal SG consisting of is output. The feedback D / A converter 14 supplies the subtraction circuit 4 with fixed pattern noise AFPN obtained by analog conversion of the digital fixed pattern noise FPN stored in the frame memory 13. The output signal SG of the subtraction circuit 4 is amplified by the amplification circuit 5 at a predetermined amplification factor. The amplification factor of the amplifier circuit 5 is set so that the amplitude (variation range) of the signal component SG obtained by removing the fixed pattern noise AFPN from the imaging signal SN is the full range of the input dynamic range of the A / D converter 6. The output signal of the amplifier circuit 5 is converted into a digital signal DG by the A / D converter 6.

The output signal DG of the A / D converter 6 includes laterally drawn noise in which the DC signal level fluctuates randomly at every horizontal scanning period.
The average value calculation circuit 15 calculates a pixel average value AV for each pixel in the horizontal effective pixel range for each horizontal scanning period. The offset calculation circuit 16 calculates an offset correction amount FS by subtracting the pixel average value AV calculated by the average value calculation circuit 15 from a predetermined fixed level reference value RF. The reference value RF is set to the center value of the dynamic range of the output signal DG of the A / D converter 6 (the center value of the range of values that the signal can take). When the output signal DG of the A / D converter 6 is, for example, an 8-bit digital signal (takes a value within a range from 0 to 255), the reference value RF is set to 127 that is the above-described center value. .

The offset correction circuit 8 adds the offset correction amount FS calculated by the offset calculation circuit 16 to the output signal of the delay circuit 7 and outputs an offset-corrected signal FG.
The delay circuit 7 calculates the offset correction amount FS of the horizontal scanning line by the average value calculation circuit 15 and the offset calculation circuit 16 for the output signal DG of the A / D converter 6 for a predetermined period, that is, for the horizontal scanning line. Is delayed by a delay time corresponding to the time until completion of, for example, one horizontal scanning period.
The output D / A converter 19 converts the output signal FG of the offset correction circuit 8 from analog to output from the output terminal 9. The signal output from the output terminal 9 is sent to the display unit 25, and an image corresponding to the signal is displayed.

  The details of the operation of suppressing the horizontal noise included in the output signal DG of the A / D converter 6 will be described with reference to FIGS. 2 to 4 show a case where the number of pixels in the horizontal direction is 100 and the number of lines in the vertical direction is 4, each of which has a horizontal pixel position on the horizontal axis and the circuit in FIG. 1 on the vertical axis. The level of the signal appearing at each position is taken. The output signal DG of the A / D converter 6 is an 8-bit digital signal, for example, and the signal level takes a value in the range from 0 to 255. In order to make the description easy to understand, the output signal DG of the A / D converter 6 will be described on the assumption that the output signal DG includes horizontally drawn noises having different sizes.

  FIG. 2 shows the level of the signal DGm (m = 1, 2, 3, 4) for each line output from the A / D converter 6. The signals DG1 and DG3 of the line L1 and the line L3 are larger than the central value 127, and are visually recognized as white stripes on the screen. The signals DG2 and DG4 of the line L2 and the line L4 are smaller than the central value 127, and are visually recognized as slightly black streaks on the screen. As described above, the signals of the lines L1 to L4 (outputs of the A / D converter 6) are represented by DG1 to DG4, and the signal of the line m is generally represented by DGm. However, the description applies to any line. In this case, it may be simply expressed as “DG” or “DGm”. The same applies to other symbols.

FIG. 3 shows an average value AVm (m = 1, 2, 3, 4) for each line (every horizontal scanning period) of the output signal of the average value calculation circuit 15, that is, the signal output from the A / D converter 6. Indicates. This average value is obtained by averaging signal values for all pixels in the horizontal effective pixel range in each line.
In FIG. 3, it is assumed that the average value AV1 of the line L1 is 150, the average value AV2 of the line L2 is 30, the average value AV3 of the line L3 is 220, and the average value AV4 of the line L4 is 60.

  3 also subtracts the pixel average values AV1 to AV4 calculated by the average value calculation circuit 15 from the output signal FS of the offset calculation circuit 6, that is, the reference value RF 127 supplied from the reference value generation means 23. The offset correction amount FSm (m = 1, 2, 3, 4) calculated in the above is shown. The offset correction amount FS1 of the line L1 is 127-150 = -23, the offset correction amount FS2 of the line L2 is 127-30 = + 97, the offset correction amount FS3 of the line L3 is 127-220 = -93, and the offset correction amount of the line L4 It is assumed that FS4 is obtained as 127-60 = + 67.

FIG. 4 shows an operation in which the offset correction circuit 8 performs offset correction by adding the offset correction amount FSm (m = 1, 2, 3, 4) calculated by the offset calculation circuit 16 to the output signal of the delay circuit 7. . The output signal of the delay circuit 7 is the A / D converter of FIG. 2 for a delay time corresponding to the time until the calculation of the offset correction amount FSm of the horizontal scanning line in the average value calculation circuit 15 and the offset calculation circuit 16 is completed. 6 output signal DG is delayed.
The signal on line L1 is offset corrected by FS1 = −23, the signal on line L2 is offset corrected by FS2 = + 97, the signal on line L3 is offset corrected by FS3 = −93, and the signal on line L4 is only FS4 = + 67. Offset correction is performed. In this way, the signals DG1 to DG8 of each line are offset-corrected so that the average of the signals FG1 to FG8 after the offset correction of each line matches the center value CD of the dynamic range. By performing the processing in this way, the horizontal noise in the lines L1 to L4 is suppressed.

  FIG. 5 to FIG. 5 show examples of signal patterns different from the above-described examples (examples shown in FIG. 2 to FIG. 4) regarding the details of the operation of suppressing the horizontal noise included in the output signal DG of the A / D converter 6. Reference is made to FIG. 5 to 14 illustrate an example in which the number of pixels in the horizontal direction is 100 and the number of lines in the vertical direction is 8. In FIGS. 6 and 8 to 13, the horizontal axis represents the horizontal pixel, and the vertical axis represents the signal level.

FIG. 5 shows the position of each part of the subject in association with the line number and pixel number of the image obtained by imaging. FIG. 7 shows an image corresponding to a subject (image displayed when noise is not superimposed and offset correction according to the present invention is not applied). FIG. 14 schematically shows a display image.
The output signal DG of the A / D converter 6 is, for example, an 8-bit digital signal and takes a value in the range from 0 to 255.

  As shown in FIG. 5, the imaging element 3 captures a high-temperature object (target object) in the lower right corner of the low-temperature background 101a, that is, (in the corresponding image, the right half 101b of the lines L5 to L8. When included in the range, the output signal from the image sensor 3 is, for example, the upper four lines out of all eight lines, that is, the lines L1 to L4 are signals corresponding to the low temperature portion 101a, and the lower four lines, that is, the lines L5 L8 outputs a signal that changes from the low temperature portion 101a to the high temperature portion 101b, however, the output signal of the image sensor 3 includes laterally drawn noise.

If the signal obtained by imaging the subject shown in FIG. 5 contains no noise (not superimposed), the value of the signal obtained from the low temperature part 101a is 127, and the signal obtained from the high temperature part 101b Assume that the value is 191.
In this case, the value of the signal obtained from the lines L1 to L4 is 127 throughout the line as indicated by reference numeral 102a in FIG. 6, and the value of the signal obtained from the lines L5 to L8 is indicated by reference numeral 102b in FIG. In each line, the first to 50th pixels are 127, while the 51st to 100th pixels are 191.

FIG. 7 shows an image when such a signal (not including noise) is used for display without correction according to the present invention. In FIG. 7, the brightness of each part (region) corresponds to the temperature of the corresponding part in FIG. 5. Note that the higher the temperature, the larger the signal value (tone value), and the brighter the displayed image.
The area 103a in FIG. 7 corresponds to the background (low temperature part) 101a in FIG. 5, the brightness of the area 103a corresponds to the signal value 127, and the area 103b corresponds to the high temperature part 101b in FIG. The area 103b is brighter than the area 103a and has a brightness corresponding to the signal value 191.

  FIG. 8 shows the levels of the signals DG1 to DG4 for each line with respect to the lines L1 to L4 among all the eight lines output from the A / D converter 6. The signals DG1 and DG3 of the line L1 and the line L3 are larger than 127, which is the center value of the dynamic range, and are visually recognized as white stripes on the screen. The signals DG2 and DG4 of the line L2 and the line L4 are smaller than 127, which is the center value of the dynamic range, and are visually recognized as slightly black streaks on the screen.

FIG. 9 shows pixel average values AV1 to AV4 for each line of the output signal of the average value calculation circuit 15, that is, the output signal of the A / D converter 6. The average values AV1 to AV4 are obtained by averaging the signal values for all the pixels in the horizontal effective pixel range in each line.
In FIG. 9, the average value AV1 of the signal DG1 of the line L1 is 159, the average value AV2 of the signal DG2 of the line L2 is 103, the average value AV3 of the signal DG3 of the line L3 is 167, and the average value AV4 of the signal DG4 of the line L4 is It is assumed that 79 was obtained.

  FIG. 9 also shows that the average value calculation circuit 15 calculates the output signal FSm (m = 1, 2, 3, 4) of the offset calculation circuit 6, that is, the reference value RF 127 supplied from the reference value generation means 23. The offset correction amount FSm (m = 1, 2, 3, 4) calculated by subtracting the obtained pixel average values AV1 to AV4 is shown. The offset correction amount FS1 of the line L1 is 127-159 = −32, the offset correction amount FS2 of the line L2 is 127-103 = + 24, the offset correction amount FS3 of the line L3 is 127-167 = −40, and the offset correction amount of the line L4 It is assumed that FS4 is obtained as 127−79 = + 48.

FIG. 9 also shows an operation of offset correction by the offset correction circuit 8 by adding the offset correction amount FS calculated by the offset calculation circuit 16 to the output signal of the delay circuit 7. The output signal of the delay circuit 7 is output from the A / D converter 6 for a delay time corresponding to the time until the calculation of the offset correction amount FS of the horizontal scanning line in the average value calculation circuit 15 and the offset calculation circuit 16 is completed. The signal DG is delayed.
The signal DG1 of the line L1 is offset-corrected by −32 which is the offset correction amount FS1, the signal DG2 of the line L2 is offset-corrected by +24 which is the offset correction amount FS2, and the signal DG3 of the line L3 is the offset correction amount FS3−. The offset correction is performed by 40, and the signal DG4 of the line L4 is offset corrected by +48 which is the offset correction amount FS4. That is, the signals DG1 to DG4 of the lines L1 to L4 are offset-corrected so that the average of the corrected signals FG1 to FG4 of each line coincides with the center value CD of the dynamic range 127. By performing the processing in this manner, horizontal noise in the lines L1 to L4 is suppressed, and the offset correction circuit 8 outputs a signal in which horizontal noise is suppressed.

FIG. 10 shows the levels of the signals DG5 to DG8 for each of the lines L5 to L8 out of all eight lines of the output signal DG of the A / D converter 6. Of the lines L5 and L7, the signals DG5a and DG7a of the low temperature part 101a are larger than the center value CD of the dynamic range 127, and are visually recognized as white stripes on the screen. The signals DG6a and DG8a of the low temperature part 101a of the line L6 and the line L8 are smaller than the central value CD of the dynamic range 127, and are visually recognized as a slightly black streak on the screen.
The signals DG5b to DG8b of the high-temperature portion 101b of the lines L5 to L8 are larger than 127, which is the center value CD of the dynamic range, and are visually recognized as horizontal streaky noise that is generally whitish on the screen.

FIG. 11 shows pixel average values AV5 to AV8 for each line of the output signal of the average value calculation circuit 15, that is, the output signal of the A / D converter 6. The average values AV5 to AV8 are obtained by averaging the signal values for all the pixels in the horizontal effective pixel range in each line.
It is assumed that the average value of the signal DG5 of the line L5 is 191, the average value of the signal DG6 of the line L6 is 135, the average value of the signal DG7 of the line L7 is 199, and the average value of the signal DG8 of the line L8 is 111.

  11 also subtracts the pixel average values AV5 to AV8 calculated by the average value calculation circuit 15 from the output signal FS of the offset calculation circuit 6, that is, the reference value RF 127 supplied from the reference value generation means 23. The calculated offset correction amounts FS5 to FS8 are shown. The offset correction amount FS5 of the line L5 is 127-191 = -64, the offset correction amount FS6 of the line L6 is 127-135 = -8, the offset correction amount FS7 of the line L7 is 127-199 = -72, and the offset correction of the line L8 It is assumed that the quantity FS8 is found to be 127−111 = + 16.

FIG. 11 also shows an operation of offset correction by the offset correction circuit 8 by adding the offset correction amount FS calculated by the offset calculation circuit 16 to the output signal of the delay circuit 7. The output signal of the delay circuit 7 is output from the A / D converter 6 for a delay time corresponding to the time until the calculation of the offset correction amount FS of the horizontal scanning line in the average value calculation circuit 15 and the offset calculation circuit 16 is completed. The signal DG is delayed.
The signal DG5 of the line L5 is offset-corrected by −64 which is the offset correction amount FS5, the signal DG6 of the line L6 is offset-corrected by −8 which is the offset correction amount FS6, and the signal DG7 of the line L7 is the offset correction amount FS7. The offset is corrected by −72, and the signal DG8 of the line L8 is offset corrected by +16 which is the offset correction amount FS8. As described above, the signals DG5 to DG8 of the lines L5 to L8 are offset-corrected so that the averages of the signals FG5 to FG8 of the corrected lines coincide with the center value CD of the dynamic range 127. By performing the processing in this manner, horizontal noise in the lines L5 to L8 is suppressed, and the offset correction circuit 8 outputs a signal in which horizontal noise is suppressed.

12 shows the output signals FG1 to FG4 of the offset correction circuit 8 for each of the lines L1 to L4, and FIG. 13 shows the output signals FG5 to FG8 of the offset correction circuit 8 for each of the lines L5 to L8. Show. Signals FG <b> 1 to FG <b> 8 shown in FIGS. 12 and 13 are sent from the output terminal 9 to the display unit 25. An image displayed on the display unit 25 by a signal output from the output terminal 9 is as shown in FIG.
As shown in the figure, in the area composed of the lines L1 to L4, the average brightness V3 (= 127) of each line is substantially equal to the brightness V2 (= 127) of the area 103a in FIG. Among the lines L5 to L8, the brightness V4 (= 95) of the left half area 104b is smaller than the brightness V1 (= 127) of the area 103a of FIG. 7, while the brightness V5 (= 159) is smaller than the brightness V2 (= 191) of the area 103b in FIG. The average brightness (V4 + V5) / 2 of each line of the lines L5 to L8 is equal to the average brightness V3 of the lines L1 to L4.
Further, in the region 104a composed of the lines L1 to L4 as well as the 104b and 104c composed of the lines L5 to L8, the horizontal noise is suppressed. As described above, as a result of performing the offset correction, although there is a deviation from the actual signal level in the lines L5 to L8, the horizontally drawn noise is suppressed, and a calm and easy-to-view image is obtained.

  Next, the operation of storing the digital fixed pattern noise FPN in the frame memory 13 (FPN capturing process) will be described with reference to FIG. At this time, the shutter 2 is controlled to be in a shielded state, and the update circuit 12 is controlled to be permitted to update the frame memory 13. Since the shutter 2 is in a light-shielded state, only a noise component including fixed pattern noise (FPN) is output from the image sensor 3 as an output signal.

  The subtraction circuit 4 is a fixed pattern obtained by analog-converting the digital fixed pattern noise (FPN) stored in the frame memory 13 from the imaging signal that is the output signal of the imaging device 3 by the feedback D / A converter 14. The noise AFPN is subtracted. The output signal of the subtraction circuit 4 is amplified by the amplification circuit 5 at a predetermined amplification factor and then converted into a digital signal by the A / D converter 6.

  The attenuation circuit 10 outputs a signal obtained by attenuating the output signal DG of the A / D converter 6 with a predetermined attenuation rate. The predetermined attenuation rate of the attenuation circuit 10 is set so that the amplitude of the signal output from the attenuation circuit 10 is the same as or smaller than the original signal (output of the subtraction circuit 4) in consideration of the predetermined amplification rate of the amplification circuit 5. The attenuation rate is as follows.

  The adder circuit 11 includes a fixed pattern noise FPN stored in the frame memory 13 (which may be expressed as FPN (i−1) because it is a fixed pattern noise acquired one frame before), and the attenuation circuit 10. The output signals are added and output as the latest digital fixed pattern noise (sometimes expressed as FPN (i)). (Note that as described above, the fixed pattern noise acquired one frame before is represented by the code “FPN (i−1)”, and the fixed pattern noise of the current frame is represented by the code “FPN (i)”. In the description that applies to the fixed pattern noise of the frame, it is simply expressed as “FPN”.) The digital fixed pattern noise (FPN (i−1)) added by the adder circuit 11 is output from the image sensor 3 in the subtractor 4. The fixed pattern noise AFPN subtracted from SG is equal to the signal FPN (i−1) before analog conversion.

  The latest digital fixed pattern noise (FPN (i)) output from the adder circuit 11 is stored in the frame memory 13 via the update circuit 12. The signal stored in the frame memory 13 is converted into the analog signal AFPN by the feedback D / A converter 14 in the next frame period, and then supplied to the subtraction circuit 4 and also to the addition circuit 11.

  The timing generation circuit 17 outputs a signal for instructing the operation timing of the shutter control circuit 18 and the update circuit 12. The shutter control circuit 18 controls the opening / closing operation of the shutter 2 based on the timing signal output from the timing generation circuit 17. The update circuit 12 updates the stored data in the frame memory 13 with the output signal of the adder circuit 11 based on the timing signal output from the timing generation circuit 17. The timing generation circuit 17 permits the update data stored in the frame memory 13 to be updated by the update circuit 12 when the shutter 2 is completely shielded.

  The timing generation circuit 17 includes digital fixed pattern noise (FPN) stored in the frame memory 13 by closing the shutter 2 in the manual mode by operating the external switch SW in the output signal of the imaging device 3 at that time. Update to a value corresponding to fixed pattern noise (FPN). Further, the timing generation circuit 17 includes digital fixed pattern noise (FPN) stored in the frame memory 13 by closing the shutter 2 every time a predetermined time elapses, in the output signal of the imaging device 3 at that time. Update to a value corresponding to fixed pattern noise (FPN).

  In the first frame period after the start of the FPN capturing process, the frame memory 13 resets the storage signal to 0 in order to newly capture digital fixed pattern noise (FPN). The feedback D / A converter 14 converts the 0 level signal output from the frame memory 13 into an analog signal and supplies it to the subtraction circuit 4. Since the subtraction circuit 4 subtracts the analog conversion signal of the 0 level signal stored in the frame memory 13 from the output signal of the image sensor 3, the output signal of the image sensor 3 is output as it is. The output signal of the subtraction circuit 4 is amplified by the amplification circuit 5 at a predetermined amplification factor and then converted into a digital signal by the A / D converter 6. At this stage, since the fixed pattern noise (FPN) output from the image sensor 3 is input to the A / D converter 6 as it is, the fixed pattern noise (FPN) larger than the input dynamic range of the A / D converter 6 is used. ) Is A / D converted only in the central part of the amplitude. Signals outside the input dynamic range are discarded without being A / D converted. At this time, the DC potential of the signal supplied to the A / D converter 6 is adjusted to be the center of the input dynamic range of the A / D converter 6.

  The attenuation circuit 10 attenuates the output signal DG of the A / D converter 6. The adder circuit 11 adds the reset 0 level signal stored in the frame memory 13 and the output signal of the attenuation circuit 10, and digital fixed pattern noise (FPN) in the first frame period after the start of the FPN capturing process. Is generated. Digital fixed pattern noise (FPN) output from the adder circuit 11 is stored in the frame memory 13 via the update circuit 12.

  In the second frame period after the start of the FPN acquisition process, the subtraction circuit 4 uses the digital fixed pattern noise (FPN) in the first frame period after the start of the FPN acquisition process, from the fixed pattern noise (FPN) output from the image sensor 3. ) Subtract analog conversion signal. For example, when the amplification factor of the amplifier circuit 5 and the attenuation factor of the attenuation circuit 10 are matched, a portion that is not A / D converted by the A / D converter 6 and is discarded in the first frame period after the start of the FPN capturing process All or a part of the signal becomes an output signal of the subtracting circuit 4.

  The output signal of the subtraction circuit 4 is amplified by the amplification circuit 5 at a predetermined amplification factor and then converted into a digital signal by the A / D converter 6. Only the central portion of the amplitude of the output signal of the image sensor 3 is A / D converted. Signals outside the input dynamic range are discarded without being A / D converted. At this time, the DC potential of the signal supplied to the A / D converter 6 is adjusted to be the center of the input dynamic range of the A / D converter 6.

  The attenuation circuit 10 attenuates the output signal DG of the A / D converter 6. The adder circuit 11 adds the digital fixed pattern noise (FPN) of the first frame period after the start of the FPN acquisition process stored in the frame memory 13 and the output signal of the attenuation circuit 10, and after the FPN acquisition process starts. Digital fixed pattern noise (FPN) of the second frame period is generated. The digital fixed pattern noise (FPN) output from the adder circuit 11 in the second frame period after the start of the FPN capturing process is stored in the frame memory 13 via the update circuit 12.

  As described above, the fixed pattern noise (FPN) written in the output signal of the image sensor 3 is digitized by performing the writing of the fixed pattern noise (FPN) for a predetermined frame period with the shutter 2 in the shielding state. Noise (FPN) can be taken into the frame memory 13. The number of frames necessary for capturing the fixed pattern noise (FPN) is obtained from the amplitude of the fixed pattern noise (FPN) included in the output signal of the image sensor 3, the amplification factor of the amplification circuit 5, the attenuation factor of the attenuation circuit 10, and the like. . The timing generation circuit 17 controls the shutter control circuit 18 and the update circuit 12 so that the fixed pattern noise (FPN) capturing process is continued for a predetermined number of frame periods. The difference between the digital fixed pattern noise (FPN) stored in the frame memory 13 in the predetermined number of frame periods and the fixed pattern noise (FPN) included in the output signal of the image sensor 3 can be ignored. This corresponds to the time required.

  Further, it may be determined that the capture is completed when the output signal DG of the A / D converter 6 becomes smaller than a predetermined value. The timing generation circuit 17 confirms that the output signal DG of the A / D converter 6 has become smaller than a predetermined value for one frame period or more, and then shifts to the standard operation state (normal imaging state), so that the shutter control circuit 18 and the update circuit 12 are controlled.

  Compared with the case where the image signal output from the image sensor 3 is directly converted into a digital signal by digitally converting the signal obtained by subtracting the fixed pattern noise (FPN) from the image signal output from the image sensor 3, the dynamic range of the digital converter Can be used effectively.

  For example, even if the A / D converter 6 and the feedback D / A converter 14 have the same bit precision, the output signal DG of the A / D converter 6 has a difference in performance such as amplitude and linearity of the output signal. Is added as it is to the digital fixed pattern noise FPN (the value during the capture process, that is, the signal that has not yet converged) as it is, depending on the characteristic difference between the A / D converter 6 and the feedback D / A converter 14, Diverges and becomes difficult to converge. Therefore, after the amplitude is attenuated by a predetermined attenuation rate by the attenuation circuit 10, it is added to the digital fixed pattern noise (FPN). As a result, the characteristic difference between the A / D converter 6 and the feedback D / A converter 14 can be absorbed, and the digital fixed pattern noise (FPN) can be converged and taken in by preventing signal divergence.

  As described above, according to the first embodiment, the signal level for each horizontal scanning line is stably maintained by calculating the level fluctuation for each horizontal scanning line and correcting the signal level of the imaging signal to a constant value. However, it is possible to suppress the horizontal noise on the screen.

Embodiment 2. FIG.
FIG. 15 is a block diagram showing the configuration of the infrared imaging apparatus according to Embodiment 2 of the present invention. In FIG. 15, the lens 1, the shutter 2, the imaging device 3, the subtraction circuit 4, the amplification circuit 5, the A / D converter 6, the attenuation circuit 10, the addition circuit 11, the update circuit 12, the frame memory 13, the feedback D / A The converter 14, the timing generation circuit 17, the shutter control circuit 18, the delay circuit 7, the offset correction circuit 8, the output D / A converter 19, the output terminal 9, the average value calculation circuit 15, and the offset calculation circuit 16 The circuit elements having the same numbers as those in FIG. 1 described in the first embodiment are the same as those in FIG.

The infrared imaging device shown in FIG. 15 is different from the infrared imaging device of FIG. 1 in that the pixel range setting circuit 22 is provided.
The pixel range setting circuit 22 sets a pixel range (average calculation pixel range) used for calculating the average among the pixels within the horizontal effective pixel range of each line, and the pixel ( A signal indicating the start pixel) and the end position pixel (end pixel) is output.

  The average value calculation circuit 15 is a pixel for a range (average calculation pixel range) from a start pixel to an end pixel indicated by a signal input from the pixel range setting circuit 22 among the pixels in the horizontal effective pixel range of each line. An average value (line average value) AV is calculated and supplied to the offset calculation circuit 16.

  The effect of the offset correction operation performed by the offset correction circuit 8 by limiting the range of pixels used when the average value is calculated by the average value calculation circuit 15 will be described below with reference to FIGS.

16 to 39 show a case where the number of pixels in the horizontal direction is 100 and the number of lines in the vertical direction is 8. 17, 19, 21, 24, 26, 28, 31, 32, 34, 35, 37, and 38 are horizontal pixels on the horizontal axis and signal levels on the vertical axis, respectively. Have taken.
FIGS. 16, 23, and 30 show the subject in correspondence with the line number and pixel number of the image obtained by imaging the position of each part.
18, 25, and 33 show images corresponding to the subject (images displayed when noise is not superimposed and offset correction according to the present invention is not applied).
20, 22, 27, 29, 36, and 39 schematically show display images.
The output signal DG of the A / D converter 6 is an 8-bit digital signal, for example, and takes a value in the range of 0 or 255.

  As one subject (first subject), a target object (target subject) 105d composed of two portions 105b and 105c having a slight temperature difference at room temperature at the center of a low-temperature background 105a as shown in FIG. Will be described. In the example shown in the figure, the 1st to 25th pixels and the 76th to 100th pixels correspond to the background 105a, and more than the portion 105b made up of the 26th to 50th pixels of the object of interest 105d. The portion 105c composed of the 51st to 75th pixels has a slightly higher temperature.

  The signal of each line obtained by imaging such a subject is, for example, as indicated by reference numeral 105e in FIG. 17 if noise is not superimposed. As shown in FIG. 17, the signal value is 0 for the 1st to 25th pixels and the 76th to 100th pixels corresponding to the background portion 105a, and for the 26th to 50th pixels, The signal value is slightly lower than 127, the signal values of the 51st through 75th pixels are slightly higher than 127, and the average of the signal values of the 26th through 75th pixels is 127.

  When such a signal (not including noise) is used for display without correction according to the present invention, an image is as shown in FIG. In FIG. 18, regions 106a, 106b, and 106c correspond to portions 105a, 105b, and 105c in FIG. 16, respectively. It is assumed that the higher the subject temperature, the larger the signal value (tone value) and the brighter the displayed image.

  First, the pixel range setting circuit 22 sets the first pixel as the start pixel and the 100th pixel as the end pixel. That is, when the entire line is set as the average calculation pixel range, the line L1 The signals DG1 to DG8 of .about.L8 are offset corrected so that the average value of the signals FG1 to FG8 after offset correction of the entire line of each line matches the reference value RF (center value of the dynamic range of the signal) 127. Thus, an image displayed using the signal output from the output terminal 9 is as shown in FIG. In FIG. 20, areas 107a, 107b, and 107c are brighter than the corresponding areas 106a, 106b, and 106c in FIG. In particular, although the central regions 107b and 107c (corresponding to the regions 106b and 106c in FIG. 18) where the target object is present become whitish images, the horizontal noise is suppressed and a calm and easy-to-view image is obtained.

  Next, the pixel range setting circuit 22 sets the 26th pixel as a start pixel and the 75th pixel as an end pixel. That is, from the 26th pixel to the 75th pixel (pixels in the central portion) Is set as the average calculation pixel range, each line average value is calculated using 50 pixels at the center in the horizontal direction. In the case of the subject shown in FIG. 16, the 1st to 25th pixels and the 76th to 100th pixels are included in the low temperature portion 105a of the background, so these are used for calculating the line average value. First, the line average value is calculated only for the subject portions 105b and 105c at room temperature of interest.

  In the signals DG1 to DG8 of the lines L1 to L8, the average value of the signals FG1 to FG8 after the offset correction of the average calculation pixel range of each line matches the reference value RF (the center value of the dynamic range of the signal) 127. Thus, the offset correction is performed as shown in FIG. 21, and the image displayed using the signal output from the output terminal 9 is as shown in FIG. In FIG. 22, regions 108a, 108b, and 108c have the same brightness as the corresponding regions 106a, 106b, and 106c in FIG. For this reason, an image having a sense of contrast is obtained without causing the central regions 108b and 108c where the object of interest is present to be over-exposed or black-out, and laterally drawn noise is suppressed, and a calm and easy-to-view image is obtained.

  As another subject (second subject), an imaging object includes a target object (target subject) 109d composed of portions 109a and 109b having a slight temperature difference at room temperature on the right half of a low-temperature background 109a as shown in FIG. The case will be described. In the illustrated example, the 1st to 50th pixels correspond to the background 109a, and the 76th to 100th pixels of the portion 109b made up of the 51st to 75th pixels in the object of interest 109d. The portion 109c made of is slightly higher in temperature.

  The signal of each line obtained by imaging such a subject is, for example, as shown in FIG. 24 if no noise is superimposed. As indicated by reference numeral 109e in FIG. 24, the first to 50th pixels corresponding to the background portion 109a have a signal value of 0, and the 51st to 75th pixels have a signal value slightly lower than 127. The signal values of the 76th to 100th pixels are slightly higher than 127, and the average of the signal values of the 51st to 100th pixels is 127.

  When such a signal (not including noise) is used for display without correction according to the present invention, an image is as shown in FIG. In FIG. 25, regions 110a, 110b, and 110c correspond to portions 109a, 109b, and 109c in FIG. 23, respectively. It is assumed that the higher the subject temperature, the larger the signal value (tone value) and the brighter the displayed image.

  First, when the pixel range setting circuit 22 sets the average calculation pixel range from the first pixel to the 100th pixel, that is, the entire line, the signals DG1 to DG8 of the lines L1 to L8 are set to the reference value RF ( The center value of the dynamic range of the signal 127) is offset-corrected so that the average values of the signals FG1 to FG8 after the offset correction of the entire line of each line coincide with each other, as shown in FIG. An image displayed using the output signal is as shown in FIG. In FIG. 27, regions 111a, 111b, and 111c are brighter than the corresponding regions 110a, 110b, and 110c in FIG. In particular, although the right-half regions 111b and 111c (corresponding to the portions 110b and 110c in FIG. 25) with the target object become whitish images, the horizontal noise is suppressed and a calm and easy-to-view image is obtained.

  Next, when the pixel range setting circuit 22 sets the 51st pixel to the 100th pixel (right half portion pixels) as average calculation pixels, 50 pixels in the right half portions 109b and 109c in the horizontal direction are set. The average value of each line is calculated. Therefore, the low temperature portion 109a of the background is not used for the calculation of the line average value, and the line average value is calculated only for the subject portions 109b and 109c of the room temperature at which attention is paid.

In the signals DG1 to DG8 of the lines L1 to L8, the average value of the signals FG1 to FG8 after the offset correction of the average calculation pixel range of each line matches the reference value RF (the center value of the dynamic range of the signal) 127. As shown in FIG. 28, the offset is corrected as shown in FIG. 28, and the image displayed using the signals FG1 to FG8 output from the output terminal 9 is as shown in FIG.
29, areas 112a, 112b, and 112c have the same brightness as the corresponding areas 110a, 110b, and 110c in FIG. For this reason, an image having a sense of contrast is obtained without causing the right-half regions 112b and 112c where the object of interest is present to be over-exposed or black-out, and the horizontally drawn noise is suppressed, and a calm and easy-to-view image is obtained.

  Further, as another subject (third subject), an imaging object includes a target object (target subject) 113d composed of portions 113b and 113c having a slight temperature difference at room temperature in the upper right portion of a low-temperature background 113a as shown in FIG. In addition, a case will be described in which an imaging object includes a target object (target target) 113g composed of portions 113e and 113f having a slight temperature difference at room temperature in the lower left part. In the illustrated example, in the lines L1 to L4, the first to 50th pixels correspond to the background 113a, and the 76th pixel than the portion 113b composed of the 51st to 75th pixels of the object of interest 113d. The portion 113c consisting of the first through 100th pixels has a slightly higher temperature. In the lines L5 to L8, the 51st through 100th pixels correspond to the background 113a, and the 26th through 50th pixels of the target object 113g are composed of the 1st through 25th pixels. The portion 113f composed of the pixels up to is slightly higher in temperature.

A signal obtained by imaging such a subject is, for example, as shown in FIGS. 31 and 32 if no noise is superimposed. 31 shows the signal of the lines L1 to L4 by reference numeral 113h, and FIG. 32 shows the signal of the lines L5 to L8 by reference numeral 113k.
In the signals of the lines L1 to L4, as shown in FIG. 31, the first to 50th pixels corresponding to the background portion 113a have a signal value of 0, and the 51st to 75th pixels have a signal value. The value is slightly lower than 127, the signal value is slightly higher than 127 in the 76th to 100th pixels, and the average signal value of the 51st to 100th pixels is 127.
In the signals of the lines L5 to L8, as shown in FIG. 32, the signal value is 0 in the 51st through 100th pixels corresponding to the background portion 113a, and the signal in the 1st through 25th pixels The value is slightly lower than 127, the signal values of the 26th to 50th pixels are slightly higher than 127, and the average of the signal values of the 1st to 50th pixels is 127.

  FIG. 33 shows an image when such a signal (not including noise) is used for display without correction according to the present invention. In FIG. 33, regions 114a, 114b, 114c, 114e, and 114f correspond to portions 113a, 113b, 113c, 113e, and 113f in FIG. 30, respectively. It is assumed that the higher the subject temperature, the higher the signal value (tone value) and the brighter the displayed image.

  First, when the pixel range setting circuit 22 sets the average calculation pixel range from the first pixel to the 100th pixel, that is, the entire line, the signals DG1 to DG4 of the lines L1 to L4 are the reference values (signals). 34 is obtained by performing offset correction so that the average value of the signals FG1 to FG4 after offset correction of the entire line of each line is equal to 127 which is the center value of the dynamic range of FIG. The signals DG5 to DG8 of the lines L5 to L8 are offset-corrected so that the average value of the signals FG5 to FG8 after offset correction of each line matches the reference value 127 (the center value of the dynamic range of the signal). An image displayed using a signal output from the output terminal 9 is as shown in FIG. 36, regions 115a, 115b, 115c, 115e, and 115f are brighter than the corresponding regions 114a, 114b, 114c, 114e, and 114f in FIG. 33, respectively. In particular, the upper right areas 115b and 115c (corresponding to 114e and 114f in FIG. 33) and the lower left areas 115e and 115f (corresponding to 114e and 114f in FIG. 33) with the object of interest become whitish images. Noise is suppressed and a calm and easy-to-view image is obtained.

  Next, the pixel range setting circuit 22 sets the average calculation pixel range from the 51st pixel to the 100th pixel (the right half of each line) for the lines L1 to L4, and the first for the lines L5 to L8. When the average pixel range from the first pixel to the 50th pixel (the left half of each line) is set as the average calculation pixel range, each of the lines L1 to L4 is set using 50 pixels of the right half portions 113b and 113c in the horizontal direction. The average value of the lines is calculated, and for the lines L5 to L8, the average value of each line is calculated using 50 pixels of the left half portions 113e and 113f in the horizontal direction. Accordingly, the low temperature portion 113a of the background is not used for the calculation of the line average value, and the line average value is calculated only from the subject portions 113b, 113c, 113e, and 113f at room temperature of interest.

  The signals DG1 to DG4 of the lines L1 to L4 are set so that the average value of the signals FG1 to FG4 after the offset correction of the average value calculation range of each line coincides with the reference value RF (the center value of the dynamic range of the signal) 127. The offset correction is performed as shown in FIG. 37, and the signals DG5 to DG8 of the lines L5 to L8 are offset to the reference value RF (the center value of the dynamic range of the signal) 127 to the average calculation range of each line. The offset correction is performed so that the average values of the subsequent signals FG5 to FG8 coincide with each other as shown in FIG. An image displayed using a signal output from the output terminal 9 is as shown in FIG. In FIG. 39, regions 116a, 116b, 116c, 116e, and 116f have the same brightness as the corresponding regions 114a, 114b, 114c, 114e, and 114f in FIG. For this reason, an image having a contrast feeling can be obtained without causing overexposure or blackout in the upper right areas 116b and 116c and the lower left areas 116e and 116f where the object of interest is present, and the horizontal noise is suppressed and settled. An easy-to-see image is obtained.

  As described above, the case where the average calculation pixel range of each line is 50 pixels has been described. However, if the pixel range is a power of 2 such as 32 pixels or 64 pixels, the average value is calculated without performing normal division. Since the line average value can be calculated with a simple bit shift, the circuit scale can be reduced. For example, when the average calculation pixel range is 64 pixels, the average value can be calculated by truncating the last 6 bits of the total data.

  In applications in which the positions of the background portion of the subject and the portion of interest (the subject of interest) are fixed, signals indicating the start pixel and the end pixel of the average calculation pixel range set by the pixel range setting circuit 22 are fixed. Based on the above, the average value calculation circuit 15 calculates the line average value, so that particularly the image of the attention portion can be visually recognized with good contrast.

  The setting of the average calculation pixel range by the pixel range setting circuit 22 is, for example, all pixels (the whole of each line), a predetermined number of pixels in the central portion, any other specified pixel range, or an average calculation pixel for each line. Whether to set a range may be selected by a switch (not shown), or an arbitrary pixel range may be input by user interface means such as a keyboard (not shown).

Embodiment 3 FIG.
FIG. 40 is a block diagram showing the configuration of the infrared imaging apparatus according to Embodiment 3 of the present invention. In FIG. 40, a lens 1, a shutter 2, an image pickup device 3, a subtraction circuit 4, an amplification circuit 5, an A / D converter 6, an attenuation circuit 10, an addition circuit 11, an update circuit 12, a frame memory 13, and a feedback D / A. The converter 14, the timing generation circuit 17, the shutter control circuit 18, the delay circuit 7, the offset correction circuit 8, the output D / A converter 19, the output terminal 9, the average value calculation circuit 15, and the offset calculation circuit 16 Since it operates in the same manner as the circuit elements with the same numbers in FIG. 1 described in the first embodiment and exhibits the same effect, the description thereof is omitted.

  The infrared imaging apparatus shown in FIG. 40 is different from the infrared imaging apparatus of FIG. 1 in that a reference value generation unit 24 is provided instead of the reference value generation unit 23 of FIG. The illustrated reference value generation means 24 includes a mixing circuit 20 and a memory 21. 1 generates the fixed reference value RF, whereas the reference value generation unit 24 of FIG. 40 generates the reference value RFm () for each line according to the average value of the input signals. i) is changed. Here, m represents a line number. In the description applicable to any line, “m” may not be attached, and it may be simply expressed by “RF”. Further, i represents a frame number, but in the description applicable to any frame, (i) is omitted and represented by “RF” and “RFm”. The same applies to other symbols.

  The reference value generation unit 24 calculates a frame direction average value for each line, and supplies the frame direction average value to the offset calculation circuit 16 as a reference value. Here, the frame direction average value is a simple average or a weighted average value of an average value of each line in the current frame and an average value of the same line in a predetermined number (1 or more) of past frames. Is obtained.

  The memory 21 is connected so that the output of the average value calculation circuit 15 is written, and the memory 21 stores the average value calculated by the average value calculation circuit 15 for a predetermined number (represented by (M−1)) frames. The average value AV (i) calculated by the average value calculation circuit 15 for the current frame and the average values AV (i−1) and AV (i−2) for (M−1) frames stored in the memory 21. ,... AV (i−M + 1) is input to the mixing circuit 20, and the mixing circuit 20 adds them and divides them by M to obtain a reference value for the current frame.

For example, M is 32. However, it may be set to a numerical value such as 16, 32, 64, or 128, or may be another numerical value.
The memory 21 needs to have a memory capacity for storing an average value obtained for each line of M-1, for example, 31 frames, and always corresponds to the past (M-1) frames, for example, 31 frames. It is necessary to perform memory control so that pixel average values AV (i-1), AV (i-2),... AV (i-31) are stored for each line.

  In the case of obtaining the weighted average, for example, the average value AV (i) obtained for the current frame and the average value AV (i−1) obtained for each of a predetermined number of frames (1 or more) in the past. , AV (i-2),... Are given different weights.

  The offset calculating circuit 16 subtracts the pixel average value AV (i) calculated by the average value calculating circuit 15 from the line reference value RF (i) of the current frame output from the reference value generating means 24, particularly the mixing circuit 20. To calculate an offset correction amount FS (i).

The offset correction circuit 8 adds the offset correction amount FS (i) calculated by the offset calculation circuit 16 to the output signal of the delay circuit 7 and outputs an offset-corrected signal FG (i).
The delay circuit 7 is a delay time corresponding to the time until the calculation of the offset correction amount FS (i) of the horizontal scanning line is completed by the average value calculation circuit 15 and the offset calculation circuit 16, for example, one horizontal scanning period). The output signal DG of the A / D converter 6 is delayed and supplied to the offset correction circuit 8. The output D / A converter 19 converts the output signal FG (i) of the offset correction circuit 8 from analog to output from the output terminal 9.

  The details of the operation of suppressing the horizontal noise included in the output signal DG of the A / D converter 6 will be described with reference to FIGS. 5 to 7, 41 to 45, and 46 to 53. FIG. 41 to 53 show the case where the number of pixels in the horizontal direction is 100 and the number of lines in the vertical direction is 8. In FIGS. 41 to 52, the horizontal axis represents the pixel position in the horizontal direction, and the vertical axis represents The level of the signal appearing at each position in the circuit of FIG. 40 is taken.

As described in connection with the first embodiment, FIG. 5 shows the subject, and the position of each part thereof is associated with the line number and pixel number of the image obtained by imaging the subject. FIG. 7 shows an image corresponding to a subject (image displayed when noise is not superimposed and offset correction according to the present invention is not applied). FIG. 53 schematically shows a display image.
The output signal DG of the A / D converter 6 is an 8-bit digital signal, for example, and the signal level takes a value in the range from 0 to 255.

  As shown in FIG. 5, the imaging device 3 captures an object having a high-temperature object in the lower right corner portion of the low-temperature background 101a, that is, in the right half 101b of the lines L5 to L8 (in the corresponding image). In this case, the output signal from the image sensor 3 is, for example, the upper four lines out of all eight lines, that is, the lines L1 to L4 are signals corresponding to the low temperature part 101a, and the lower four lines, that is, the lines L5 to L8 are low temperature. The signal which changes from the part 101a to the high temperature part 101b is output. However, the output signal of the image sensor 3 includes horizontal noise.

  If the signal obtained from the subject shown in FIG. 5 does not include noise, the value of the signal DG obtained from the low temperature part 101a is 127, and the value of the signal DG obtained from the high temperature part 101b is 191. The value of the signal DG obtained from the lines L1 to L4 is as shown in FIG.

  FIG. 7 shows an image when such a signal (not including noise) DG is used for display without performing correction according to the present invention.

  However, actually, the output signal of the image sensor 3 includes noise, and the magnitude of the noise may change from frame to frame. Therefore, in the third embodiment, for each line, the line average value for the past (M−1) frames is stored in the memory 21, and each of the past (M−1) frames stored in the memory 21 is stored. The average of the line average values AVm (i− (M−1)) to AVm (i−1) and the line average value AVm (i) calculated for the current frame is obtained by the mixing circuit 20, and this is calculated for the current frame. A line reference value (RFm (i)) for the line is output.

  Hereinafter, for the sake of simplicity, the calculation operation of the line reference value RFm in the mixing circuit 20 and the memory 21 in the case of M = 4 will be described with reference to numerical examples with reference to FIGS.

FIG. 41 is a diagram for explaining the calculation of the line reference value RF1 for the line L1. For example, a case where the line reference value RF1 is obtained by averaging four data of the average value of the line L1 for four fields will be described. The average value calculation circuit 15 supplies 159 to the mixing circuit 20 as the average value AV1 (i) of the line L1 in the current field F (i). The memory 21 stores 113 as the average value AV1 (i-3) of the line L1 of the third previous field F (i-3) and the average value of the line L1 of the second previous field F (i-2). AV1 (i-2) 141 is stored, and 95 is stored as the average value AV1 (i-1) of the line L1 of the previous field F (i-1).
The mixing circuit 20 receives 113 from the memory 21 as the average value AV1 (i-3) of the line L1 of the field F (i-3) and the average value AV1 (i− of the line L1 of the field F (i-2). 2) and 95 as the average value AV1 (i-1) of the line L1 of the field F (i-1) are read out, and the line L1 of the field F (i) supplied from the average value calculation circuit 15 is read out. As shown in FIG. 41, an average AA1 of four data (average value of line L1 for four fields) together with 159 which is an average value AV1 (i) of
AA1 (i) = (113 + 141 + 95 + 159) / 4 = 127
And AA1 (i) = 127 is supplied to the offset calculation circuit 16 as the line reference value RF1 (i) for the line L1.

FIG. 42 is a diagram for explaining the calculation of the line reference value RF2 for the line L2. For example, a case where the line reference value RF2 is obtained by averaging four data of the average value of the line L2 for four fields will be described. The average value calculation circuit 15 supplies 103 to the mixing circuit 20 as the average value AV2 (i) of the line L2 of the current field F (i). The memory 21 stores 151 as the average value AV2 (i-3) of the line L2 in the third previous field F (i-3) and the average value of the line L2 in the second previous field F (i-2). 79 as AV2 (i-2) and 175 as the average value AV2 (i-1) of the line L2 of the previous field F (i-1) are stored.
The mixing circuit 20 receives from the memory 21 151 as the average value AV2 (i-3) of the line L2 of the field F (i-3) and the average value AV2 (i− of the line L2 of the field F (i−2)). 2) and 175 as the average value AV2 (i-1) of the line L2 of the field F (i-1) are read out, and the line L2 of the field F (i) supplied from the average value calculation circuit 15 is read out. As shown in FIG. 42, the average AA2 (i) of four data (average value of the line L2 for four fields) together with the average value AV2 (i) 103 of
AA2 (i) = (151 + 79 + 175 + 103) / 4 = 127
And AA2 (i) = 127 is supplied to the offset calculation circuit 16 as the line reference value RF2 (i) for the line L2.

FIG. 43 is a diagram for explaining the calculation of the line reference value RF3 for the line L3. For example, a case where the line reference value RF3 is obtained by averaging four data of the average value of the line L3 for four fields will be described. The average value calculation circuit 15 supplies 167 to the mixing circuit 20 as the average value AV3 (i) of the line L3 in the current field F (i). The memory 21 stores 87 as the average value AV3 (i-3) of the line L3 of the third previous field F (i-3) and the average value of the line L3 of the second previous field F (i-2). 137 as AV3 (i-2) and 117 as the average value AV3 (i-1) of the line L3 of the previous field F (i-1) are stored.
The mixing circuit 20 reads 87 from the memory 21 as the average value AV3 (i-3) of the line L3 of the field F (i-3) and the average value AV3 (i− of the line L3 of the field F (i-2). 2) and 117 as the average value AV3 (i-1) of the line L3 of the field F (i-1) are read out, and the line L3 of the field F (i) supplied from the average value calculation circuit 15 is read out. As shown in FIG. 43, the average AA3 (i) of four data (the average value of the line L3 for four fields) is combined with 167 which is the average value AV3 (i) of
AA3 (i) = (87 + 137 + 117 + 167) / 4 = 127
And AA3 (i) = 127 is supplied to the offset calculation circuit 16 as the line reference value RF3 (i) for the line L3.

FIG. 44 is a diagram for explaining the calculation of the line reference value RF4 for the line L4. For example, a case where the line reference value RF4 is obtained by averaging four data of the average value of the line L4 for four fields will be described. 79 is supplied from the average value calculation circuit 15 to the mixing circuit 20 as the average value AV4 (i) of the line L4 in the current field F (i). The memory 21 stores 175 as the average value AV4 (i-3) of the line L4 in the third previous field F (i-3) and the average value of the line L4 in the second previous field F (i-2). 99 as AV3 (i-2) and 155 as the average value AV2 (i-1) of the line L4 of the previous field F (i-1) are stored.
The mixing circuit 20 reads from the memory 21 175 as the average value AV4 (i-3) of the line L4 in the field F (i-3) and the average value AV4 (i−) of the line L4 in the field F (i-2). 99) as 2) and 155 as the average value AV4 (i-1) of the line L4 of the field F (i-1) are read out, and the line L4 of the field F (i) supplied from the average value calculation circuit 15 is read. 44, the average AA4 (i) of the four data (the average value of the line L4 for four fields) together with 79, which is the average value AV4 (i) of
AA4 (i) = (175 + 99 + 155 + 79) / 4 = 127
And AA4 (i) = 127 is supplied to the offset calculation circuit 16 as the line reference value RF4 (i) for the line L4.

  In the above example, the averages AA1 (i) to AA4 (i) of a plurality of frames of the lines L1 to L4 coincide with each other and are all 127, but this is a coincidence, and in fact, these are They do not always coincide with each other and do not necessarily become 127, but tend to be close to 127, and tend to be closer to 127 as the number of frames M is increased. Here, for the sake of simplicity, the description is made assuming that it matches 127.

45, the pixel average value AVm (i) of the current frame calculated by the average value calculation circuit 15 is subtracted from the line reference value RFm (i) of each line generated by the mixing circuit 20 as the output signal FS of the offset calculation circuit 6. The offset correction amount FSm calculated in the above is shown.
The offset correction amount FS1 (i) of the line L1 is from the reference value RF1 (i) = 127 of the line L1 and the average value AV1 (i) = 159 of the line L1 FS1 (i) = 127-159 = −32
Accordingly, the offset correction amount FS2 (i) of the line L2 is changed from the reference value RF2 (i) = 127 of the line L2 and the average value AV2 (i) = 103 of the line L2 to FS2 (i) = 127−103 = + 24.
Accordingly, the offset correction amount FS3 (i) of the line L3 is changed from the reference value RF3 (i) = 127 of the line L3 and the average value AV3 (i) = 167 of the line L3 to FS3 (i) = 127-167 = −40.
Accordingly, the offset correction amount FS4 (i) of the line L4 is changed from the reference value RF4 (i) = 127 of the line L4 and the average value AV4 (i) = 79 of the line L4 to FS4 (i) = 127−79 = + 48.
Respectively.

FIG. 45 also shows an operation of offset correction by the offset correction circuit 8 by adding the offset correction amount FS calculated by the offset calculation circuit 16 to the output signal of the delay circuit 7. The output signal of the delay circuit 7 is A / D converted in FIG. 2 by a delay time corresponding to the time until the calculation of the offset correction amount FS of the horizontal scanning line is completed by the average value calculation circuit 15 and the offset calculation circuit 16. The output signal DG of the device 6 is delayed.
The signal of the line L1 is offset-corrected by −32 which is the offset correction amount FS1 (i), the signal of the line L2 is offset-corrected by +24 which is the offset correction amount FS2 (i), and the signal of the line L3 is offset correction amount FS3. (I) is offset-corrected by −40, and the signal of the line L4 is offset-corrected by +48 which is an offset correction amount FS4 (i). That is, in this way, the signals of the lines L1 to L4 are offset corrected so that the averages of the signals FG1 (i) to FG4 (i) after the offset correction of each line coincide with the line reference value RFm 127. Is done. In this way, the horizontally drawn noise in the lines L1 to L4 corresponding to the low temperature portion is suppressed.

FIG. 46 is a diagram for explaining the calculation of the line reference value RF5 for the line L5. For example, the case where the line reference value RF5 is obtained by averaging four data of the average value of the line L5 for four fields will be described. The average value calculation circuit 15 supplies 191 to the mixing circuit 20 as the average value AV5 (i) of the line L5 of the current field F (i). The memory 21 stores 127 as the average value AV5 (i-3) of the line L5 in the third previous field F (i-3) and the average value of the line L5 in the second previous field F (i-2). 221 as AV5 (i-2) and 97 as the average value AV5 (i-1) of the line L5 of the previous field F (i-1) are stored.
The mixing circuit 20 reads from the memory 21 127 as the average value AV5 (i-3) of the line L5 in the field F (i-3) and the average value AV5 (i−) of the line L5 in the field F (i−2). 2) and 97 as the average value AV5 (i-1) of the line L5 of the field F (i-1) are read out, and the line L5 of the field F (i) supplied from the average value calculation circuit 15 is read out. The average value AA5 (i) of four data (average value of the line L5 for four fields) together with 191 that is the average value AV5 (i) of AA5 (i) = (127 + 221 + 97 + 191) / 4 as shown in FIG. = 159, and AA5 (i) = 159 is supplied to the offset calculation circuit 16 as the line reference value RF5 of the line L5.

FIG. 47 is a diagram for explaining the calculation of the line reference value RF6 for the line L6. For example, the case where the line reference value RF6 is obtained by averaging four data of the average value of the line L6 for four fields will be described. The average value calculation circuit 15 supplies 135 to the mixing circuit 20 as the average value AV6 (i) of the line L6 in the current field F (i). The memory 21 stores 183 as the average value AV6 (i-3) of the line L6 in the third previous field F (i-3) and the average value of the line L6 in the second previous field F (i-2). 105 as AV7 (i-2) and 213 as the average value AV8 (i-1) of the line L6 of the previous field F (i-1) are stored.
The mixing circuit 20 reads from the memory 21 183 as the average value AV6 (i-3) of the line L6 in the field F (i-3) and the average value AV6 (i−) of the line L6 in the field F (i-2). 2) and 213 as the average value AV6 (i-1) of the line L6 of the field F (i-1) are read out, and the line L6 of the field F (i) supplied from the average value calculation circuit 15 is read out. As shown in FIG. 47, the average AA6 (i) of the four data (average value of the line L6 for four fields) together with the average value AV6 (i) of 135 is shown in FIG.
AA6 (i) = (183 + 105 + 213 + 135) / 4 = 159
And AA6 (i) = 159 is supplied to the offset calculation circuit 16 as the line reference value RF6 (i) of the line L6.

FIG. 48 is a diagram for explaining the calculation of the line reference value RF7 for the line L7. For example, the case where the line reference value RF7 is obtained by averaging four data of the average value of the line L7 for four fields will be described. The average value calculation circuit 15 supplies 199 to the mixing circuit 20 as the average value of the line L7 in the current field F (i). The memory 21 stores 148 as the average value AV7 (i-3) of the line L7 in the third previous field F (i-3) and the average value of the line L7 in the second previous field F (i-2). 170 as AV7 (i-2) and 119 as the average value AV7 (i-1) of the line L7 of the previous field F (i-1) are stored.
The mixing circuit 20 reads from the memory 21 148 as the average value AV7 (i-3) of the line L7 in the field F (i-3) and the average value AV7 (i−) of the line L7 in the field F (i-2). 2) and 119 as the average value AV7 (i-1) of the line L7 of the field F (i-1) are read out, and the line L7 of the field F (i) supplied from the average value calculation circuit 15 is read out. As shown in FIG. 48, the average AA7 (i) of four data (average value of the line L7 for four fields) together with 199 which is the average value AV7 (i) of
AA7 (i) = (148 + 170 + 119 + 199) / 4 = 159
And AA7 (i) = 159 is supplied to the offset calculation circuit 16 as the line reference value RF7 (i) of the line L7.

FIG. 49 is a diagram for explaining the calculation of the line reference value RF8 for the line L8. For example, a case where the line reference value RF8 is obtained by averaging four data of the average value of the line L8 for four fields will be described. The average value calculation circuit 15 supplies 111 to the mixing circuit 20 as the average value of the line L8 in the current field F (i). The memory 21 stores 135 as the average value AV8 (i-3) of the line L8 in the third previous field F (i-3) and the average value of the line L8 in the second previous field F (i-2). 207 as AV8 (i-2) and 183 as the average value AV8 (i-1) of the line L8 of the previous field F (i-1) are stored.
The mixing circuit 20 reads 135 from the memory 21 as the average value AV8 (i-3) of the line L8 in the field F (i-3) and the average value AV8 (i−) of the line L8 in the field F (i-2). 2) and 183 as the average value AV8 (i-1) of the line L8 of the field F (i-1) are read out and the line L8 of the field F (i) supplied from the average value calculation circuit 15 is read out. As shown in FIG. 49, the average AV8 (i) of four data (the average value of the line L8 for four fields) together with the average value AV8 (i) 111 of
AA8 (i) = (135 + 207 + 183 + 111) / 4 = 159
And AA8 (i) = 159 is supplied to the offset calculation circuit 16 as the line reference value RF8 (i) of the line L8.

  In the above example, the averages AA5 (i) to AA8 (i) of the plurality of frames of the lines L5 to L8 coincide with each other and are all 159, but this is a coincidence, and in fact, these are They do not necessarily match and do not always become 159, but tend to be close to 159, and tend to be closer to 159 as the number of frames M is increased. Here, for the sake of simplicity, the description is made assuming that it matches 159.

In FIG. 50, the pixel average value AVm (i) of the current frame calculated by the average value calculation circuit 15 is subtracted from the line reference value RFm (i) of each line generated by the mixing circuit 20 as the output signal FS of the offset calculation circuit 6. The offset correction amount FSm calculated in the above is shown.
The offset correction amount FS5 (i) of the line L5 is from the reference value RF5 (i) = 159 of the line L5 and the average value AV5 (i) = 191 of the line L5 FS5 (i) = 159-191 = −32
Accordingly, the offset correction amount FS6 (i) of the line L6 is changed from the reference value RF6 (i) = 159 of the line L6 and the average value AV6 (i) = 135 of the line L6 to FS6 (i) = 159−135 = + 24.
Therefore, the offset correction amount FS7 (i) of the line L7 is from the reference value RF7 (i) = 159 of the line L7 and the average value AV7 (i) = 199 of the line L7, FS7 (i) = 159-199 = −40, The offset correction amount FS8 (i) of the line L8 is from the reference value RF8 (i) = 159 of the line L8 and the average value AV8 (i) = 111 of the line L8 FS8 (i) = 159−111 = + 48
Respectively.

FIG. 50 also shows an operation of offset correction by the offset correction circuit 8 by adding the offset correction amount FS calculated by the offset calculation circuit 16 to the output signal of the delay circuit 7. The output signal of the delay circuit 7 is the A / D converter shown in FIG. 40 for a delay time corresponding to the time until the calculation of the offset correction amount FS of the horizontal scanning line in the average value calculation circuit 15 and the offset calculation circuit 16 is completed. 6 output signal DG is delayed.
The signal DG5 of the line L5 is offset-corrected by −32 which is the offset correction amount FS5 (i), the signal DG6 of the line L6 is offset-corrected by +24 which is the offset correction amount FS6 (i), and the signal DG7 of the line L7 is offset. The offset correction is performed by −40 which is the correction amount FS7 (i), and the signal DG8 of the line L8 is offset corrected by +48 which is the offset correction amount FS8 (i). That is, in this way, the signals DG5 to DG8 of the lines L5 to L8 are offset so that the average of the corrected signals FG5 (i) to FG8 (i) of each line matches 159 which is each line reference value RFm. It is corrected. In this way, the horizontal noise in the lines L5 to L8 that change from the low temperature portion to the high temperature portion is suppressed.

In this way, the offset correction circuit 8 outputs the signal FGm in which the horizontal noise is suppressed.
51 shows signals FG1 (i) to FG4 (i) obtained by offset-correcting the signals DG1 (i) to DG4 (i) for the lines L1 to L4, and FIG. 52 shows the signals for the lines L5 to L8. Signals FG5 (i) to FG8 (i) obtained by offset correction of DG5 (i) to DG8 (i) are shown. This signal is sent from the output terminal 9 to the display unit 25. An image displayed using a signal output from the output terminal 9 is as shown in FIG.
In FIG. 53, the brightness of each part (region) corresponds to the temperature of the corresponding part in FIG. Note that the higher the temperature, the larger the signal value (tone value), and the brighter the displayed image.
An area 117a in FIG. 53 corresponds to the background (low temperature part) 101a in FIG. 5, the brightness of the area 117a corresponds to the signal value 127, and the area 117b corresponds to the high temperature part 101b in FIG. The region 117b is brighter than the region 117a and has brightness corresponding to the signal value 191.
In this way, the display image does not deviate from the actual signal level, and the horizontal noise is suppressed, and a calm and easy-to-view image is obtained. In addition, an image signal corresponding to the subject in FIG.

  As described above, according to the third embodiment, the signal level for each horizontal scanning line is made constant by calculating the level fluctuation for each horizontal scanning line and correcting the signal level of the imaging signal to the line reference value. It is stable, and horizontal noise can be suppressed even on the screen.

  Further, since the line reference value is generated from the pixel average value AV of the line and the average value of the past frame, the signal level difference for each line due to the difference in the temperature level of the subject can be eliminated.

Embodiment 4 FIG.
FIG. 54 is a block diagram showing an infrared imaging apparatus according to Embodiment 4 of the present invention. The infrared imaging apparatus of FIG. 54 is generally the same as the apparatus of FIG. 40, but the connection and operation of the mixing circuit 20 and the memory 21 constituting the reference value generating means 24 are different.
54, the memory 21 stores the reference value calculated by the mixing circuit 20, and the reference value stored in the memory 21 is supplied to the mixing circuit 20.
In the apparatus of FIG. 54, the average value for each line of the current frame and the average value of the average value for the same line in a predetermined number of past frames are not obtained as in the apparatus of FIG. By calculating the weighted average of the average value calculated for each line in the average value calculation means 15 and the reference value obtained for the previous frame for the same line and stored in the memory 21 by the mixing circuit 20, Determine the reference value for the current frame for the current frame.
For example, the weighting coefficient for the average value AVi obtained in the current frame is 1 / N (N is a value greater than 1), and the weighting coefficient for the reference value RF (i−1) in the previous frame is {1- (1 / N)}. in this case,
The reference value RFi of the current frame is
RFi = AVi * (1 / N) + RF (i-1) * (1- (1 / N))
It is calculated by.
The reference value RFi calculated by the mixing circuit 20 for each frame is written into the memory 21 and read from the memory 21 in the next frame period to be mixed as “reference value of previous frame (RF (i−1)”. It is supplied to the circuit 20.
Here, N is a positive integer, and is preferably a power of 2, for example, 16, 32, 64, or 128.

  Since the reference value for the previous frame is obtained by repeating the same processing in the past, the calculation of the reference value by such a method takes into consideration the average value of all the past frames and is close to the present. This is a method of obtaining an average by assigning a greater weight to the object, and can be said to be a kind of method of obtaining an average in the frame direction.

  54, the memory 21 needs to have a memory capacity for storing a line reference value corresponding to each line.

  The configuration of FIG. 54 has the advantage that the same effect as the configuration of FIG. 40 can be obtained and the capacity of the memory 21 can be smaller than that of the configuration of FIG.

It is a block diagram which shows the structure of the infrared imaging device which concerns on Embodiment 1 of this invention. FIG. 2 is a diagram showing the level of a signal DGm (m = 1, 2, 3, 4) for each line output from the A / D converter 6 of FIG. 1. The pixel average value AVm (m = 1, 2, 3, 4) for each line (each horizontal scanning cycle) of the output signal of the average value calculation circuit 15 of FIG. 1, that is, the signal output from the A / D converter 6 is obtained. FIG. 1 is a diagram illustrating an operation of performing offset correction by adding the offset correction amount FSm (m = 1, 2, 3, 4) calculated by the offset calculation circuit 16 to the output signal of the delay circuit 7 in the offset correction circuit 8 of FIG. It is. It is a figure which shows an example of a to-be-photographed object, and matches the position of each part with the line number and pixel number of the image obtained by imaging. FIG. 6 is a diagram illustrating signal values obtained from lines L1 to L4 and L5 to L8 when a signal obtained by imaging the subject illustrated in FIG. 5 does not include noise. It is a figure which shows the image at the time of displaying, without performing the correction | amendment by this invention using the signal (it shall not contain noise) shown by FIG. FIG. 6 is a diagram showing signals DG1 to DG4 of lines L1 to L4 output from the A / D converter 6 when the subject shown in FIG. 5 is imaged. FIG. 6 is a diagram illustrating pixel average values AV1 to AV4 for each line of an output signal of the average value calculation circuit 15 when the subject shown in FIG. 5 is imaged, that is, an output signal of the A / D converter 6; FIG. 6 is a diagram illustrating signals DG5 to DG8 of lines L5 to L8 output from the A / D converter 6 when the subject illustrated in FIG. 5 is imaged. The pixel average values AV5 to AV8 for each line of the output signal of the average value calculation circuit 15 when the subject shown in FIG. 5 is imaged, that is, the output signal of the A / D converter 6 are shown. FIG. 6 is a diagram showing output signals FG1 to FG4 of an offset correction circuit 8 for each of lines L1 to L4 when the subject shown in FIG. 5 is imaged. FIG. 6 is a diagram showing output signals FG5 to FG8 of the offset correction circuit 8 for each of lines L5 to L8 when the subject shown in FIG. 5 is imaged. 14 is a diagram showing an image displayed on the display unit 25 when signals FG1 to FG8 shown in FIGS. 12 and 13 are sent from the output terminal 9 to the display unit 25. FIG. It is a block diagram which shows the structure of the infrared imaging device which concerns on Embodiment 2 of this invention. It is a figure which shows an example of a to-be-photographed object, and matches the position of each part with the line number and pixel number of the image obtained by imaging. It is a figure which shows the signal obtained from the lines L1-L8 when the signal of each line obtained by imaging the to-be-photographed object shown in FIG. 16 does not contain noise. It is a figure which shows the image at the time of displaying, without performing the correction | amendment by this invention using the signal (it shall not contain noise) shown by FIG. When the pixel range setting circuit 22 has set the first pixel as the start pixel and the 100th pixel as the end pixel, when the subject shown in FIG. 16 is imaged, the offset of FIG. It is a figure which shows signals FG1-FG8 output from the correction circuit 8. It is a figure which shows the image displayed using the signals FG1-FG8 shown by FIG. When the pixel range setting circuit 22 has set the 26th pixel as the start pixel and the 75th pixel as the end pixel, when the subject shown in FIG. 16 is imaged, the offset of FIG. It is a figure which shows signals FG1-FG8 output from the correction circuit 8. It is a figure which shows the image displayed using the signals FG1-FG8 shown by FIG. It is a figure which shows the position of each part of another example of a subject in association with the line number and pixel number of an image obtained by imaging. It is a figure which shows the signal obtained from the lines L1-L8 when the signal of each line obtained by imaging the to-be-photographed object shown in FIG. 23 does not contain noise. It is a figure which shows the image at the time of displaying, without performing the correction | amendment by this invention using the signal (it shall not contain noise) shown by FIG. When the pixel range setting circuit 22 has set the first pixel as the start pixel and the 100th pixel as the end pixel, when the subject shown in FIG. It is a figure which shows signals FG1-FG8 output from the correction circuit 8. It is a figure which shows the image displayed using the signals FG1-FG8 shown by FIG. When the pixel range setting circuit 22 has set the 51st pixel as the start pixel and the 100th pixel as the end pixel, when the subject shown in FIG. 23 is imaged, the offset of FIG. It is a figure which shows signals FG1-FG8 output from the correction circuit 8. It is a figure which shows the image displayed using signal FG1-FG8 shown by FIG. It is a figure which shows the position of each part in another example of a subject in association with the line number and pixel number of an image obtained by imaging. It is a figure which shows the signal obtained from line L1-L4 when the signal of each line obtained by imaging the to-be-photographed object shown in FIG. 30 does not contain noise. It is a figure which shows the signal obtained from the lines L5-L8 when the signal of each line obtained by imaging the to-be-photographed object shown in FIG. 30 does not contain noise. It is a figure which shows the image at the time of displaying, without performing the correction | amendment by this invention using the signal (it shall not contain noise) shown by FIG. 31, FIG. When the pixel range setting circuit 22 has set the first pixel as a start pixel and the 100th pixel as an end pixel, when the subject shown in FIG. 30 is imaged, the offset of FIG. It is a figure which shows signals FG1-FG4 about the lines L1-L4 output from the correction circuit 8. FIG. When the pixel range setting circuit 22 has set the first pixel as a start pixel and the 100th pixel as an end pixel, when the subject shown in FIG. 30 is imaged, the offset of FIG. It is a figure which shows signals FG5-FG8 about the lines L5-L8 output from the correction circuit 8. FIG. It is a figure which shows the image displayed using signal FG1-FG8 shown by FIG. 34, FIG. When the pixel range setting circuit 22 has set the 51st pixel as the start pixel and the 100th pixel as the end pixel, when the subject shown in FIG. 30 is imaged, the offset of FIG. It is a figure which shows signals FG1-FG4 about the lines L1-L4 output from the correction circuit 8. FIG. When the pixel range setting circuit 22 has set the first pixel as a start pixel and the 50th pixel as an end pixel, when the subject shown in FIG. 30 is imaged, the offset of FIG. It is a figure which shows signals FG5-FG8 about the lines L5-L8 output from the correction circuit 8. FIG. It is a figure which shows the image displayed using signal FG1-FG8 shown by FIG. 37, FIG. It is a block diagram which shows the structure of the infrared imaging device which concerns on Embodiment 3 of this invention. It is a figure which shows average value AV1 (i-3) -AV1 (i) about line L1 from the field before three to the present field, and line reference value RF1 (i) defined based on these. It is a figure which shows average value AV2 (i-3) -AV2 (i) about line L2 from the field before 3 to the present field, and line reference value RF2 (i) defined based on these. It is a figure which shows average value AV3 (i-3) -AV3 (i) about line L3 from the field before 3 to the present field, and line reference value RF3 (i) defined based on these. It is a figure which shows average value AV4 (i-3) -AV4 (i) about line L3 from the field before 3 to the present field, and line reference value RF4 (i) defined based on these. The pixel average value AV1 (i) of the current frame calculated by the average value calculation circuit 15 from the line reference values RF1 (i) to RF4 (i) of each line generated by the mixing circuit 20 as the output signal FS of the offset calculation circuit 6 It is a figure which shows offset correction amount FS1 (i)-FS4 (i) calculated by subtracting AV4 (i). It is a figure which shows average value AV5 (i-3) -AV5 (i) about line L5 from the field before 3 to the present field, and line reference value RF5 (i) defined based on these. It is a figure which shows average value AV6 (i-3)-AV6 (i) about line L6 from the field before 3 to the present field, and line reference value RF6 (i) defined based on these. It is a figure which shows average value AV7 (i-3) -AV7 (i) about line L7 from the field before three to the present field, and line reference value RF7 (i) defined based on these. It is a figure which shows average value AV8 (i-3) -AV8 (i) about line L8 from the field before 3 to the present field, and line reference value RF8 (i) defined based on these. The average pixel value AV5 (i) of the current frame calculated by the average value calculation circuit 15 from the line reference values RF5 (i) to RF8 (i) of each line generated by the mixing circuit 20 as the output signal FS of the offset calculation circuit 6. It is a figure which shows offset correction amount FS5 (i)-FS8 (i) calculated by subtracting AV8 (i). It is a figure which shows signal FG1 (i)-FG4 (i) obtained by carrying out offset correction of signals DG1 (i) -DG4 (i) about lines L1-L4. It is a figure which shows signal FG5 (i)-FG8 (i) obtained by carrying out offset correction | amendment of signal DG5 (i) -DG8 (i) about lines L5-L8. FIG. 53 is a diagram showing an image displayed using signals FG1 (i) to FG8 (i) shown in FIGS. 51 and 52. It is a block diagram which shows the structure of the infrared imaging device which concerns on Embodiment 4 of this invention.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 Lens, 2 Shutter, 3 Image pick-up element, 4 Subtraction circuit, 5 Amplification circuit, 6 A / D converter, 7 Delay circuit, 8 Offset correction circuit, 9 Output terminal, 10 Attenuation circuit, 11 Addition circuit, 12 Update circuit, 13 frame memory, 14 feedback D / A converter, 15 average value calculation circuit, 16 offset calculation circuit, 17 timing generation circuit, 18 shutter control circuit, 19 output D / A converter, 20 mixing circuit, 21 memory

Claims (10)

Imaging means having sensitivity in a predetermined wavelength range;
In an infrared imaging device including an imaging unit that forms an image of a predetermined wavelength region component on an imaging surface of the imaging unit,
Digital conversion means for converting an imaging signal output from the imaging means into a digital signal and outputting a digital imaging signal;
Line average value calculating means for calculating an average of horizontal effective pixel portions for each horizontal scanning period of the digital imaging signal output from the digital conversion means and outputting a line average value;
For each line, a line reference value calculation means for calculating a frame direction average of the line average value and outputting as a line reference value;
Offset calculating means for calculating an offset value for each horizontal scanning period from a level difference obtained by subtracting the line average value output from the line average value calculating means from the line reference value output from the line reference value calculating means; ,
Delay means for delaying the digital imaging signal output from the digital conversion means by a time for processing in the average value calculation means and the offset calculation means;
An infrared imaging apparatus comprising: an offset correction unit that performs level correction by subtracting the offset value calculated by the offset calculation unit from the digital imaging signal delayed by the delay unit. For each line, the line reference value calculation means includes a memory that stores line average values for past frames, line average values for past frames stored in the memory, and line average values calculated for the current frame. The infrared imaging apparatus according to claim 1 , further comprising a mixing unit that calculates an average and outputs the average as a line reference value for the line of the current frame. The line reference value calculating means mixes the line average value of the current frame output from the line average value calculating means for each line with the line reference value of the line of the previous frame at a predetermined ratio, and mixes the line reference value of the current frame. The infrared imaging device according to claim 1 , wherein a value is newly calculated. The mixed value of the line average value of the current frame and the line reference value of the previous frame in the line reference value calculation means is:
Calculated by adding a value obtained by multiplying the line average value of the current frame by 1 / N (where N is a value greater than 1) and a value obtained by multiplying the line reference value of the previous frame by {1− (1 / N)}. The infrared imaging device according to claim 3 .   The infrared imaging apparatus according to claim 1, wherein the line average value calculating unit calculates an average value for pixels within a predetermined average calculation pixel range within a horizontal effective pixel range. 6. The infrared imaging device according to claim 5 , wherein the average value calculation pixel range is located in a horizontal central portion within a horizontal effective pixel range. 6. The infrared imaging apparatus according to claim 5 , wherein the average calculated pixel range is a pixel range set corresponding to the position of the subject of interest. 8. The infrared imaging apparatus according to claim 7 , wherein the average calculated pixel range is a plurality of pixel ranges set in units of horizontal scanning lines corresponding to positions of a plurality of target subjects to be noticed. The infrared imaging device according to claim 5 , wherein the average calculation pixel range includes power-of-two pixels.   The infrared imaging device according to claim 1, wherein the predetermined wavelength band component is a wavelength band of about 8 to 14 micrometers.

Images