JPH0691879B2 - Image processing method in electronic endoscope apparatus - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to an image processing method in an electronic endoscope apparatus for imaging in a frame sequential method.

[Conventional technology]

The field-sequential method that sequentially obtains color images corresponding to illumination light of each color from a two-dimensional image sensor such as a CCD is effective when a plurality of two-dimensional image sensors cannot be mounted. It is being applied to endoscopic devices.

Such an electronic endoscope apparatus simultaneously converts color images sequentially captured according to illumination light via red (R), green (G), and blue (B) field memories, and outputs color images as color images. It is configured to play on a TV.

[Problems to be Solved by the Invention]

However, in an electronic endoscope apparatus that captures images in the frame-sequential method, it is necessary to make each field data of R, G, and B the same field in order to prevent line flicker. Therefore, in the past, odd field and even field were used. Only one of the field data is read out from the image sensor, or the two-pixel mixed data is added and read out at once from the image sensor.

Therefore, the resolution in the vertical direction of the reproduction screen is half the resolution originally possessed by the image sensor.

In addition, a general electronic endoscope system uses a general TV as an image sensor.
1/2 inch or 1/3 inch optical size full frame CC
D (500 (H) x 400 (V) = 200,000 pixels) is used, but the CCD applicable to a small-diameter endoscope such as a bronchoscope is a CCD for local observation with a size of about 2 mm square or less. The number of pixels is as small as tens of thousands of pixels.

Therefore, when the CCD is operated in the current system, the screen size of the reproduced image is determined by the number of pixels, and the 2 mm square CCD has a small screen size, which is a practical problem.

The present invention has been made in view of such circumstances, and provides an image processing method in an electronic endoscope apparatus capable of improving the resolution in the vertical direction and quadrupling the screen size in area ratio. The purpose is to provide.

Further, when the shape of the image sensor is small and the number of pixels is small, when the image is displayed on the monitor, the texture of the reproduced image is inevitably coarse, and defects such as blurring of the outline occur. The present invention is also aimed at solving such drawbacks, and an object thereof is to provide an image processing method in an electronic endoscope apparatus capable of obtaining a reproduced image in which a contour line is fine and clear.

Further, when an edge enhancement signal for enhancing the edge of a reproduced image is formed from frame data, noise is superimposed on the edge enhancement signal, resulting in line flicker and fixed pattern noise. . The present invention is
It is also an object of the present invention to provide an image processing method in an electronic endoscope apparatus capable of emphasizing an edge of a reproduced image without generating line flicker or fixed pattern noise for the purpose of solving such drawbacks.

[Means for Solving the Problems]

In order to achieve the above-mentioned object, the present invention obtains a color signal of a frame sequential method from an image pickup device arranged at a distal end portion of an endoscope, A / D-converts the frame sequential color signal, and sequentially stores in a memory for each color. Along with, in the image processing method in the electronic endoscope apparatus, which is converted into a simultaneous color signal by synthesizing the frame-sequential color signals read from each memory, and a color image is reproduced by the simultaneous color signal, The field sequential color signal is obtained by transferring the odd field data and the even field data of the image pickup device within one field period, and the horizontal field is obtained from the odd field data and the even field data for each horizontal scanning line. At least vertically of the reproduced image obtained from the frame data by alternately writing the data of The edge emphasis signal for emphasizing the edge in the opposite direction is formed from the frame data, and only a portion of the edge emphasis signal near the edge portion of the reproduced image is extracted and added to the frame data, and then the frame data for each color Are combined to obtain the simultaneous color signal.

Further, the frequency of the read clock for reading the color signal from each of the memories is set to half the frequency of the write clock for writing the color signal to each of the memories.

Further, the frame data read out from the memory of any color among the memories for each color is D / A converted, and the amplitude of 1H delay frame data delayed by one horizontal scanning period is doubled. Non-delayed frame data that is only D / A converted and not delayed, and after D / A conversion 2
The signal obtained by subtracting the 2H delay frame data delayed by the horizontal scanning period is used as the contour enhancement signal, and a portion only near the edge portion of the reproduced image of the contour enhancement signal is extracted.
The data is read out from the memory for each color and added to the 1H delay frame data of each color delayed for one horizontal scanning period after D / A conversion, and then the 1H delay frame data of each color is combined to perform the simultaneous processing. It is characterized in that a formula color signal is obtained.

[Operation] According to the present invention, both the odd field data and the even field data of the image sensor are read out within one field period, and one horizontal field is read out from the odd field data and the even field data in the memory of the corresponding color channel. Data for each scanning line is alternately used as write frame data, and a contour emphasis signal for enhancing at least a vertical edge of a reproduced image obtained from the frame data is formed from the frame data, and the reproduced image of the contour emphasized signal is formed. Since only the portion near the edge portion of is extracted and added to the frame data and then the frame data is combined to obtain the simultaneous color signal, the resolution in the vertical direction should be twice that of the conventional method. And at least the vertical contour lines are fine and clear. An accurate sharp reproduced image can be obtained, and line flicker and fixed pattern noise are not generated on the reproduced screen.

Further, by making the read clock frequency 1/2 times the write clock frequency to the memory, the screen size can be made 4 times as large as the area ratio.

Further, only the vicinity of the edge portion of the contour emphasis signal obtained by subtracting the non-delayed frame data and the 2H delayed frame data from the doubled 1H delayed frame data is added to the 1H delayed frame data for each color. Therefore, at least the contour line in the vertical direction is strongly emphasized with a simple arithmetic circuit structure, and at the same time, a reproduced image without line flicker and fixed pattern noise can be obtained.

〔Example〕

Hereinafter, preferred embodiments of an image processing method in an electronic endoscope apparatus according to the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a block diagram showing an embodiment of an electronic endoscope apparatus to which the method of the present invention is applied. Light from an illumination lamp 10 passes through a near infrared ray removing filter 12, a color filter disk 14 and a light guide 16. The subject 18 is illuminated from the tip of the endoscope. That is, the color filter disk 14 has a red filter, a green filter, and a blue filter, respectively, and is rotated by the motor 20 at a predetermined rotation speed (for example, 20 rps). As a result, the light from the illumination lamp 10 sequentially changes through the rotating color filter disk 14 in a cycle of 1/60 seconds, that is, a red (that changes in synchronization with the cycle of the VD pulse shown in FIG. 2A). The illumination light of each color of R), green (G), and blue (B) is obtained (FIG. 2 (B)), and the subject 18 is passed through the light guide 16.
Added to.

An image pickup lens 22 provided at the tip of the endoscope picks up an image of a subject 18 illuminated by red, green, and blue illumination light and forms an image on a light receiving portion of a CCD 24.

The CCD 24 has, for example, a size of several mm square, and is, for example, an interline CCD having 141 pixels in the vertical direction and 100 pixels in the horizontal direction.
Then, it is driven by the timing pulse applied by the synchronizing signal generating circuit 26 through the CCD driving circuit 28. That is,
The image information which is incident on the light receiving part of the CCD 24 and photoelectrically converted is the second
As shown in FIG. 6C, the odd field data and the even field data of the CCD 24 are sequentially interline-transferred within one field period, and the odd field data and the even field data are A /
It is added to the D converter 32.

Here, the interlaced scanning will be described in more detail below.
As shown in the figure, within one field period (16.6 mS), data for horizontal scanning lines ˜ to configure odd field O are transferred first, and then data for horizontal scanning lines ′ to ′ to configure even field E are transferred. Is transferred.

In addition, the transfer time of data for one horizontal scanning line is set to H /
It is set to 2 (≈31.75 μS), and the data transfer time for one frame is shortened to approximately 4.61 mS (≈31.8 μS × 141 + 2H). In addition, 2H in the figure shows the period of the protrusion to the vertical transfer path. In this way, by increasing the transfer rate from the CCD24 and reading it out, the irradiation time of light is
The output voltage gain of is obtained.

Further, the horizontal transfer (write) frequency is set to 2f _sc ≈7.16 MHz.

The A / D converter 32 converts the field sequential color signals (analog signals of the odd field O and the even field E) corresponding to each input illumination light into digital signals and outputs the digital signals to the changeover switch 34.

The change-over switch 34 changes over the input frame-sequential color signal to any one of the three memories 40, 42 and 44 and outputs it in synchronization with a timing pulse of one field period applied from the synchronizing signal generating circuit 26. Movable contact piece 34E
Are sequentially switched to the contacts 34A, 34B, 34C, 34A ... And the frame sequential color signals corresponding to the respective illumination lights are sequentially output to the memories 40 to 44.

The three memories 40 to 44 temporarily store frame data (odd field data and even field data) for each color, and writing and reading of frame data are controlled by a control signal from the memory drive circuit 48.

Next, the writing / reading control of the memory will be described.

As shown in FIGS. 2 (C) and 3, the frame sequential color signals of the odd field O and the even field E are sequentially written in the corresponding memories, and at the same time, as shown in FIG. 2 (D).
Also, as shown in FIGS. 4A to 4C, R, G, and B color signals are read out and converted into the simultaneous system.

Further, as shown in FIG. 4C, when writing the odd field data and the even field data, the data of the first horizontal scanning line of the odd field O, the data of the first horizontal scanning line of the even field E, The data of the second horizontal scanning line of the odd field O, the data of the 70th horizontal scanning line of the even field E, and the data of the 71th horizontal scanning line of the odd field O are in this order.
The data of the odd field O and the data of the even field E are alternately written for each horizontal scanning line to complete one frame data. Of course, during the field period on the TV monitor (16.6 mS, 262.5 lines), the reading period of each horizontal scanning line is 1H (63.5 μS), and the reading period of one frame data is about 8.95 mS (≈63.5 μS ×
141). Further, the read clock frequency f _{sc at} this time is 3.58 MHz, which is half that at the time of writing.

Returning to FIG. 1 again, the frame data read from the memories 40, 42, 44 are respectively set as R _OH , G _OH , B _OH, and each of the frame data R _OH , G _OH , B _OH is horizontally scanned. The frame data read with a delay is set as R _1H , G _1H , B _1H ,
The frame data read out by further delaying the frame data G _1H by one horizontal scanning period is referred to as G _2H .

The frame data R _1H read from the memory 40 is input to the D / A converter 50, the frame data G _OH read from the memory 42 is input to the D / A converter 52, and the frame data G _1H is D / A. The frame data G _2H is input to the converter 54 and the D / A converter 56, respectively. Further, the frame data B _1H read from the memory 44 is input to the D / A converter 58, and each D / A converter 50, 52, 5
Converted to analog signal at 4,56,58. The input frame data of each D / A converter 50, 52, 54, 56, 58 is a digital signal, and the output signal is an analog signal. Since the contents are the same, R _1H , G _OH representing each frame data
The same symbols are used for symbols such as.

Output signal G _OH of D / A converter 52, Output signal G _{1H of} D / A converter 54
The output signal G _2H of the D / A converter 56 is input to the vertical detail forming circuit 60 that forms the contour emphasis signal G _E that emphasizes the vertical contour of the reproduced image. In the vertical detail forming circuit 60, the contour emphasis signal G _E is formed by performing the arithmetic processing of G _E = 2 × G _1H − (G _OH + G _2H ) ... (1) from the respective input signals.

FIG. 5 shows an example of the process of forming the contour emphasis signal G _E.

In Figure 5, as the frame data G _OH shown in Fig. (A), at a concentration of 0 level until time t _1, becomes a signal indicating the dark green with steep rises at time t _1, the time t ₄
It is assumed that the signal keeps the same density up to and then returns to the density 0 level with a sharp fall at time t ₄ .

Since the frame data G _1H is a signal obtained by delaying the frame data G _OH by 1H, as shown in FIG. 7B, it rises at time t ₂ after 1H from time t ₁ and goes from time t ₄ to 1H.
The signal will fall at time t ₅ after a lapse of time.

Similarly, the frame data G _2H rises at time t ₃ which is 1H after time t ₂ as shown in FIG.
the falls at the time t ₆ after 1H period from t _5.

In the vertical detail forming circuit 60, the above three pieces of frame data G _OH , H _1H , G _2H are arithmetically processed according to the equation (1) and output, so that the output edge enhancement signal G _E is as shown in), the frame data G _1H rise immediately before the 1H time (from time t ₁ to time t ₂₎ indicates a negative amplitude, the frame data G 1H period from the rise of _1H (time t ₂ from time t Up to ₃ ) shows positive amplitude. Similarly, the 1H period immediately before the fall of the frame data G _1H (time t ₄
To the time t ₅ )
1H period (from time t ₅ to time t ₆ ) shows negative amplitude.

Further, the output frame data G _OH of the D / A converter 52 and the output frame data G _2H of the D / A converter 56 are input to a difference calculation circuit 61 that forms a difference signal of the input signals. The output signal D of the difference calculation circuit 61 becomes D = G _OH −G _2H , and in the example of FIG. 5, a portion (+/−) having a positive / negative level fluctuation of the contour emphasis signal G _E as shown in FIG. It becomes a pulse signal that generates a pulse voltage of positive or negative polarity in 1H period before and after the edge portion of the frame data G _1H .

The output signal D of the difference calculation circuit 61 is input to a waveform shaping circuit 63 that outputs a pulse voltage of positive polarity when a pulse input of positive or negative polarity is applied. In the case of FIG. 5, the output signal G _A of the waveform shaping circuit 63 is, as shown in FIG. 5F, a pulse voltage having a positive polarity for each period of 1H before and after the edge portion of the frame data G _1H. It becomes a pulse signal to be generated.

The contour emphasis signal G _E which is the output signal of the vertical detail forming circuit 60 is input to one input terminal of the adder 70 via the switch 65 which is on / off controlled by the output signal G _A of the waveform shaping circuit 63. It

On the other hand, the output frame data R _1H of the D / A converter 50 and the output frame data G _1H of the D / A converter 54 are input to a known horizontal detail forming circuit 62, and the horizontal direction of the reproduced image is known by a known method. A contour emphasis signal H _E for emphasizing the contour is formed and input to the other input terminal of the adder 70. The contour enhancement signal G _E is input to one input terminal of the adder 70,
At 70, the contour enhancement signal H _E is added to output a total contour enhancement signal T _E in the horizontal and vertical directions.

The contour emphasis signal T _E is input to one input terminal of each of the adders 64, 66 and 68, and the output of the D / A converter 50 is input to each of the other input terminals of the adders 64, 66 and 68. Frame data R _1H , D / A
Output frame data G _IH transducer 54, the output frame data B _IH of the D / A converter 58 is input.

In each adder 64, 66, 68, the input frame data R _1H , G _1H , B _1H and the contour emphasis signal T _E are added,
Frame data R _'1H, G' of the vertical and horizontal edge is emphasized _1H, is output as B _'IH. If shown, ignoring a contour enhancement signal H _E the waveform of the frame data G _'IH As an example, becomes as FIG. 5 (G), 1 horizontal scanning line immediately before being colored green indicates a negative green component , The next horizontal scanning line is colored green more intensely than the average green level. That is, since the green component is completely removed immediately before being colored green, and immediately after being colored dark green, the contour of the green color in the vertical direction is reproduced extremely vividly.

The same operation is performed at the trailing edge of the frame data _G'1H , and one horizontal scanning line at the trailing edge shows a dark green color, and the green component of the one horizontal scanning line immediately below it is completely decolorized, resulting in a strong green color. Is displayed.

This also applies to the other frame data R ′ _1H and B ′ _{1H. Since the} equal amounts G _E are added to the red and blue signals, respectively, the contour portion is emphasized and a sharp image is formed, but There is no color.

At this time, the contour emphasis signal G _E only outputs pulse voltages of positive and negative polarities before and after the edge portion of the frame data G _1H ,
Other times, for example, time t easy to ride noise because it is 0 level signal at between ₃ and time t _4, as frame data edge enhancement signal G _E riding of this noise when the noise is gone riding G _When added to _1H , the noise is superimposed on the signal of the high level portion of the frame data G _1H and appears as flicker noise or fixed pattern noise on the reproduction screen.

In the present embodiment, the contour emphasis signal G _E is gated by the output signal G _A of the waveform shaping circuit 63. Therefore, in the example of FIG. 5 described above, the period 1H before and after the edge portion of the frame data G _1H is excluded. Since the contour emphasizing signal G _E is not input to the adder 70, the noise component riding on the contour emphasizing signal G _E is not superimposed on the frame data G _1H , and therefore flicker noise or fixed pattern noise on the reproduction screen is Does not occur.

Output frame data R _'1H, G' of each of the adders 64, 66, 68 _IH,
B ′ _1H is input to the encoder 72, and is combined with the NTSC system color video signal by the encoder 72 and output as a composite color video signal.

In this way, since one frame data / one field period data is obtained from the CCD24, the vertical resolution can be doubled, and this one frame data can be converted into an odd field (70 lines), If you combine frames in an even field (70 lines), you get an image of 141 lines (double the image in the vertical direction), and 2 in the horizontal direction.
By reading the data written at f _sc (7.16 MHz) at f _sc (3.58 MHz), the image becomes twice as large, and the screen size can be made four times as large as the area ratio. Furthermore, since the data for each horizontal scanning line is alternately constructed from the odd field data and the even field data to form the frame data, and the contour emphasis is performed, the density of the horizontal scanning lines for emphasizing the contour is high in the vertical direction. Fine-tuned contour enhancement is possible.

In this embodiment, the odd field data is first transferred from the CCD 24 and then the even field data is transferred, but the odd field data and the even field data may be transferred alternately.
Further, the vertical detail forming circuit 60 forms the edge enhancement signal G _E by using the frame data G _OH , G _1H , and G _2H representing the green color signal. It is used in the highest wavelength band of, and therefore has a good S / N ratio and a large resolution, and even if an edge enhancement signal is formed using frame data representing other red or blue color signals. good. In addition, horizontal detail formation circuit
62 does not necessarily have to be used.

〔The invention's effect〕

As described above, according to the image processing method in the electronic endoscope apparatus according to the present invention, the resolution in the vertical direction can be improved, and the screen size can be quadrupled in area ratio. It is possible to apply a small image pickup device to an electronic endoscope, and it is possible to obtain a reproduced image in which at least a vertical contour line is fine and sharp.

Furthermore, a good quality reproduced image can be obtained without the occurrence of flicker noise or fixed pattern noise on the reproduced image.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of an electronic endoscope apparatus to which the method of the present invention is applied, and FIGS. 2A to 2D are timings of respective parts used for explaining FIG. A chart, FIG. 3 is a timing chart showing details of data transfer by interlaced scanning of a CCD, and FIGS. 4 (A) to 4 (C) are used for explaining writing of data to the memory of FIG. 1, respectively. Timing chart, FIG. 5 is an example of a waveform diagram showing the operation of the vertical detail forming circuit of FIG. 24 …… CCD, 26 …… synchronous signal generation circuit, 28 …… CCD drive circuit, 34 …… changeover switch, 40,42,44 …… memory, 48 …… memory drive circuit, 60 …… vertical detail forming circuit, 61 …… Difference calculation circuit, 62 …… Horizontal detail forming circuit, 63 …… Wave shaping circuit, 64,66 …… (Adder), 68,70 …… (Adder), 72 …… Encoder.

Claims (3)

[Claims] 1. A color signal of a frame-sequential system is obtained from an image pickup device arranged at the tip of an endoscope, and the frame-sequential color signal is converted into an A / D
In the electronic device, which is converted and sequentially stored in the memory for each color, is converted into a simultaneous color signal by synthesizing the frame sequential color signals read from each memory, and a color image is reproduced by the simultaneous color signal. In the image processing method in an endoscope apparatus, the field sequential color signal is obtained by transferring the odd field data and the even field data of the image sensor within one field period, and the odd field data and the field data are stored in each memory. Data for each horizontal scanning line is alternately written from even-numbered field data to form frame data, and a contour enhancement signal for enhancing at least a vertical edge of a reproduced image obtained from the frame data is formed from the frame data. Only near the edge of the reproduced image of the edge enhancement signal Is extracted and added to the frame data, and then the frame data for each color is combined to obtain the simultaneous color signal. 2. The electronic device according to claim 1, wherein a frequency of a read clock for reading the color signal from each memory is set to a half of a frequency of a write clock for writing the color signal to each memory. Image processing method in an endoscope apparatus. 3. The frame data read from a memory of any color among the memories of the respective colors is D / A converted and delayed by one horizontal scanning period, and the amplitude of 1H delayed frame data is doubled. The contour enhancement signal is a signal obtained by subtracting the non-delayed frame data read from the memory and D / A converted but not delayed, and the 2H delay frame data delayed by 2 horizontal scanning periods after D / A conversion, as the contour enhancement signal. Data processing is performed by extracting only a portion of the signal near the edge portion of the reproduced image, reading it from the memory for each color, and adding it to 1H delay frame data of each color delayed for one horizontal scanning period after D / A conversion. 3. The image processing in the electronic endoscope apparatus according to claim 1 or 2, wherein the 1H delay frame data for each color is combined and then the simultaneous color signal is obtained. Method.