JPS595944B2 - Special offer - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a means for extracting an image of a portion of an object having a specific light state from an image.

Conventionally, this type of technology has been able to process brightness among light conditions at high speed. That is, an electric signal obtained by an imaging device such as a TV camera, which is approximately proportional to the brightness, can be easily set to, for example, "1" for a bright area and "o" for a dark area using an analog comparator. Since the analog comparator's response time in this process is about several nanoseconds, it can respond sufficiently to video signals from a standard scanning TV camera. However, as shown in Figure 1a, when a red ball 3 is placed on a green table 2 in front of a blue wall 1, and each light and shade is represented by the density of hatch ferns, the brightness is - If we use a comparator (hereinafter referred to as processing) to extract only the balls, we will obtain a processing pattern as shown in Figure 1b. Here, the "1" portion is a ball, but not all of the balls are extracted as "1". If you change the value so as to extract all parts of the ball (change the operating level of the analog comparator), the result will be as shown in Figure 1c, and the parts other than the ball will also be in the "1" state. In the end, it is not possible to extract only the ball part as "1".
On the other hand, if we focus on the fact that the ball part is red and extract signals for each color from the imaging device and extract only the red signal, we should be able to achieve our goal. This process requires complex calculations,
Processing takes time and it is impossible to process the video signal of a standard scanning type T camera in real time. The fact that processing cannot be performed in real time has the disadvantage that after the video signal is once stored in a storage device, processing is executed for each part of the video, making the device large and complex. As can be seen from the examples described above, an object of the present invention is to obtain a pattern as shown in FIG. 1d, in which only the ball portion is extracted from an image of an object as shown in FIG. 1a.

Strictly speaking, the purpose of the present invention is to extract a part of an object having a specific color, regardless of the brightness of that part. FIG. 2 is a diagram showing the basic configuration of the device of the present invention, where 41 and 42 are image pickup devices, and 6 is a multiplier, that is, an amplifier or an attenuator at the input. device, 7 is analog comparator, 8 is C
It is an RT (cathode ray tube) display device.

In particular, 4 and 5 are imaging devices having different color sensitivity characteristics. For example, 41 is a vidicon camera with a blue filter added, and 42 is a vidicon camera with a red filter added.
It's like a camera. 4, 42 may achieve the same purpose with a single color TV camera, depending on the case.

That is, it is also possible to automatically replace the filters of the camera and capture images corresponding to each filter in a time-sharing manner.

In this case, 41 and 42 are of course integrated, but according to the example above, the signal supplied to 7 is a blue component video signal, and the signal supplied to 6 is a red component video signal. be. Of course, these are examples, and it is arbitrary what color component video signals are supplied to 6 and 7. Further, here, a display device 8 is used as a device receiving processing results for the sake of simplicity, but this device is used for convenience of explanation and is not essential to the configuration of the device of the present invention. Now, FIG. 3a is a reprint of FIG. 1a, but for the sake of simplicity, only the outline and color symbols are shown.

Again, regarding light and shade, if the density of the hatchings in Figure 1 a is high, it will be dark (2). Suppose that there is a tendency for 1. If we take the position of the part indicated by the dotted line in Figure 3a on the horizontal axis and the brightness of the corresponding position on the vertical axis, Figure 3B
, c.

Figure b corresponds to a video signal when a red filter is applied, and Figure c corresponds to a video signal when a blue filter is applied. Corresponding to Fig. 2, Fig. 3 B, c
are the output video signals of 42 and 41, respectively. If the output of the imaging device 42 is attenuated by the multiplier 6, the output of the multiplier 6 will become as shown by the broken line in FIG. 3b.
When the output signal of the multiplier 6 and the output signal of the imaging device 41 are input and they take approximately the same value, the output is "
If the analog comparator is constructed so as to be 1", its output signal will be as shown in FIG. 3d.

In other words, only the part of the red ball 3 is extracted as "111".In this way, "1"
By displaying it on the display device 8 by brightening the part ``0I'' and darkening the part ``0I'', a pattern as shown in FIG. 1d can be obtained and the desired purpose can be achieved. It is also called, and when the two input signal voltages are El and e2, for example, the output is set to "1" under the following conditions, and the output is set to "o" at other times, as shown in Figure 4. This can be realized with the following configuration.

In the figure, 9 and 10 are DC power supplies with voltages β and α, respectively, 11 and 12 are DC amplifiers with sufficiently large gains, and 1
3,135 is a NOT circuit, and 14 is an AND circuit. Now between terminal 15 and ground (hereinafter omitted as they are of the same type) e1
When a voltage is applied to the upper input terminal of 11, e1+β
voltage is applied. The amplifier 11 is now cut off and the output voltage becomes zero.

If this state is set to "0", one condition for the output of the OR circuit 14 to be "11" is met.Also, whether the lightning pressure ζ at the terminal 151fCe. If so, the amplifier 12 is cut off and the output voltage becomes 0, and in the same way, another condition under which the output of the OR circuit 14 becomes "1" is determined.

Thus, when the condition of equation (1) that satisfies both equations (2) and (3) is satisfied, the output of the OR circuit 14 becomes 11''.If such an analog comparator is used for 7 in FIG. , for example, the output from the imaging device 42 is
If the voltage attenuated by
The output of the analog comparator 7 becomes "1", and the output of the part 7 having a certain specific hue can be set to 11I.

In FIG. 3, a certain attenuation is applied to the output of the imaging device 42, and
Although only the red part is extracted, by appropriately setting the attenuation factor or amplification factor of the multiplier 6, for example, only the blue part or only the green part can be extracted by "1".
Patterns in which these processing results are displayed on the display device 8 in the same manner as described above are shown in FIGS. 1E and F, respectively.What is particularly noteworthy in FIG. There is also a change in brightness in the red ball 3, and the right side in Figure A is darker than the left side, and this is also reflected in the video signal voltage in Figures B and C.

However, according to the circuit configuration shown in FIG. 2, all red portions appear as "1" in the output of the analog comparator 7, regardless of this change in brightness. In other words, the circuit shown in Figure 2 works regardless of the brightness of the object.
It can be seen that it has a function to extract only parts with a specific color. Of course, the color of an object is made up of three primary colors, so as explained above, by simply processing the video signals from the two imaging devices filtered with red and blue filters, the colors of all the colors can be combined. It is not possible to extract only specific colors. However, it goes without saying that when the colors of objects that come into the field of view of the imaging device are limited in advance, a simple device as described above is often sufficient. Since the above explanation is somewhat qualitative, we will consider it strictly here.

All diagrams in Figure 5 show the wavelength x of light in the horizontal direction; Figure a shows the spectral sensitivity f1 of the first imaging system in the vertical direction, and Figure b shows the spectral sensitivity f1 of the second imaging system in the vertical direction. The spectral sensitivity F2 and the spectral luminance Ga of the object a to be extracted are shown in c of the figure. Among the above, spectral sensitivity indicates the overall effect such as the spectral sensitivity of the image sensor of the imaging system and the spectral transmittance of optical filters and lenses interposed between the object and the image sensor, and spectral brightness refers to the spectral brightness of the object. It is assumed to indicate the overall effect of spectral reflectance and spectral illuminance of illumination on an object. However, since the spectral illuminance of an object increases or decreases at a constant ratio over each wavelength due to the inclination of the object surface, the shadow of other objects, etc., this ratio is defined as k.

Now, when the object a is imaged by the imaging systems 1 and 2, the video signals Hl and h2 are respectively given with a constant of proportionality l.

Therefore, for object a, h is Ht due to the inclination of the surface of object a and other objects. It can be said that it is constant regardless of the influence of shadows etc.

If we rewrite this equation now, we can see that for the video signal of object a, the video signal of imaging system 2 multiplied by a constant coefficient L equal to Y matches the video signal of imaging system 1.ノ j１ １◆↓日′５－f? 4'7. View, 4
, is the imaging system 2, 6 is an awn multiplier with a constant coefficient, and 1 is a comparator for checking whether the left side and the right side are equal.

As can be seen from this, in general, a circuit for extracting the image part of object a from the image signals of two imaging systems can be represented as shown in FIG. Or it could be a multiplier where 6 is 1, 6 is 1, or 61 is 1. A multiplier with k25 and 6 with 1 (that is, k with just a conductive wire) may be used.

More generally, what about 61 and 62? It may also be a multiplier with a multiplier like .

Now, taking an example expressed by five components, it is clear that the video signal of the part of object a satisfies equation (5), but the image of an object with a spectral luminance other than that of object a It is not necessarily true that the signal does not always satisfy Equation 5.
Figure 5d shows the spectral characteristics of the area surrounded by . Now, considering a simple example in which Ga* has emission line spectral characteristics, equation (C5) is satisfied both in the case shown in Figure 5 e and in the case shown in Figure 5 f, so that (5) The formula will also be satisfied. As can be seen from this example, it is not possible to extract only the part of object a from objects containing all colors using only the determination condition of equation (5) or the circuit configuration of FIG. 2. However, among the objects that come into the field of view captured by the imaging device, the colors of objects other than object a are only 2 (as long as there is a condition that ε does not satisfy the 5 elements, as in the circuit configuration shown in Figure 2). It is possible to extract only object a using only a systematic imaging system, and there are many cases like this.Especially when the object is an object in the production process, the color of the object may be artificial. There are many colors available, and sufficient effects can be obtained with this type of circuit configuration.If the spectral sensitivity of the imaging system is F, or F2,
If, even after appropriately changing , it is still not possible to extract only a specific object a with only two imaging systems, increase the number of imaging systems to i so that holds true, and for example expand the configuration in Figure 6. Then, if the first system has two multipliers, the second system has one multiplier, and the i-th system has one multiplier, k; ▲
There is no significant effect in preventing objects other than object a from being extracted.

Of course, in such processing, equation (7) is obtained by transforming. The first one? If the configuration shown in the figure is adopted, one multiplier can be omitted.

In this case, the multiplier is omitted in the first system, but the imaging system with which spectral sensitivity is selected as the first system is completely free. In addition, in FIG. 1, the outputs of the respective multiplications by the multipliers 6, 62, ..., 6i of the second system to the i-th system are combined with the video signal of the first system and the analog signal.
Comparator, T,,73,......,Ti
are compared and the results are ANDed in the AND circuit 17, but the combination of things to be compared is arbitrary and does not necessarily have to be as shown in Figure 7 (8). ) can be easily understood by looking at the equation. In addition, when using a sufficiently large number of imaging systems, the spectral sensitivity of each system is set to a sufficiently narrow wavelength band as shown in Fig. 8, and if the center wavelengths of the systems are arranged at approximately equal intervals, it is possible to detect the object. When will the ability to extract only a become so high? Furthermore, if a multiplier with a different multiplier is provided for each imaging system and configured as shown in Fig. 9, it is possible to independently extract objects with different colors such as object b and object c in addition to object a. It goes without saying that it can be done.

In the figure, symbols are used in accordance with FIG. 7, and dedicated circuits for extracting objects A, b, and c are shown with suffixes A, b, and c, respectively. As explained above, imaging devices with multiple types of spectral sensitivities are arranged so that they each capture approximately the same field of view, and these video signals are directly input or transmitted through an amplifier with a predetermined constant amplification factor or attenuation with a predetermined constant attenuation factor. After passing through the comparator, multiple comparators are used to check whether two signals consisting of any combination of those signals have values that are close to each other, and the output of all the comparators is the input of the comparator. By configuring the apparatus to output states different from each other between two signals that correspond to approximate states and six signals that do not, it is possible to distinguish between parts of a specific color and parts that do not.

For example, a color TV camera can be used as the imaging device having multiple types of spectral sensitivities, or TV cameras using different color filters can be arranged at almost the same position so as to capture almost the same field of view. This can also be achieved by doing. Using a half mirror, each T
It goes without saying that you will get much better results if you align the optical axes of the V cameras.) The amplifier may be an amplifier using ordinary transistors or ICs, and the attenuator may be realized by a voltage dividing circuit using resistors.

Furthermore, by arranging a register and a DA converter in parallel, for example, the amplification factor and attenuation factor of these amplifiers and attenuators can be controlled by numerical values set in the register from the outside. In this way, it is possible to freely control this device using a computer or the like and extract objects having arbitrary color information. Further, the comparator can be easily realized with a circuit as described in FIG. Furthermore, logic circuits such as 13 and 14 used in this or 17 used in FIG. 7 or FIG. 9 are also extremely common. These elements are not only extremely easy to obtain, but also operate at high speed.
For example, since it can sufficiently respond to the high speed video signal of several megahertz from a standard scanning type T camera, it is possible to extract only an object image of a specific color in real time in accordance with the transmission speed of the video signal. Of course, use this processing result as is in C
When connected to an RT display device, for example, it is possible to see an image in which only parts of an object with a specific color are bright and other parts are dark, so a person can easily see the shape of an object with a specific color by looking at this display. It has the advantage of being able to know the position, posture, etc. It is not only used for simple display purposes, but also contains a huge amount of information including color and brightness information, for example, 25 times each of four colors per point on the image.
If it has a brightness change range of about 6 levels, then 10 bits/
Since point information is compressed to just 1 bit/point, this information can be used, for example, in Japanese Patent Application No. 46-97122, 47.
101882, 48-21636, 50-36758
If this information is applied to a device that recognizes the shape, position, orientation, etc. of an object as shown in the figure below, it will be possible to recognize various properties of objects of a specific color among objects with complex colors and brightness using an extremely simple device.

[Brief explanation of the drawing]

FIG. 1 is an explanatory diagram of a video to be processed and a processing output pattern of the present invention, and FIGS. 2 to 9 are explanatory diagrams of the configuration and operation of the present invention.

Claims (1)

[Claims] 1. At least two first and second photoelectric conversion means having mutually different spectral sensitivity distributions, and an output ratio of the first and second photoelectric conversion means for a specific color to be detected. It has means for outputting a signal of a value obtained by multiplying the output of the second photoelectric conversion means by a set value, and means for comparing the signal of the multiplied value and the output of the first photoelectric conversion means. A specific color extraction device characterized by: 2. A plurality of photoelectric conversion means having mutually different spectral sensitivity distributions, and a value obtained by multiplying the output of each of the photoelectric conversion means by a value proportional to the reciprocal of the output value of each of the photoelectric conversion means for a specific color to be detected. A specific color extraction device comprising means for outputting a signal and means for comparing the multiplied value signals with each other.