JPH02236689A - Picture input device - Google Patents

Abstract

PURPOSE:To obtain a picture of an excellent quality by outputting the position information of a contour part, the differential value of lightness value of two ground color parts pinching the contour part, and the lightness value of a specific reference point as picture information. CONSTITUTION:The position information of the contour part extracted by a contour extracting means 200, the differential value of the lightness value pinching the contour part of the two ground color parts, which are calculated by an arithmetic means 300, and the lightness value of the specific reference point is outputted as the picture information. Consequently on a picture processor side, which receives these pieces of information, the lightness value at the reference point position is set as an initial value, the lightness value of a picture element up to the position of the contour part is set at the same value as the initial value, and for the picture element position of the other ground color part exceeding the contour part, by adding the differential value to the set value, the lightness value of the picture element for the whole picture, namely, the picture information can be constructed. Thus the lightness value of each picture element of the ground color part out of the constructed picture information is set at a constant value, and the excellent picture can be obtained.

Description

[Detailed Description of the Invention] [Field of Application in Business A] The present invention relates to a human-powered image device that converts an optical image from a subject into an electrical image signal and outputs it, and is unaffected by uneven lighting. The present invention relates to an image input device capable of outputting correct brightness and compressing an image signal into elements.

[Prior Art] Conventionally, image input devices generally perform photoelectric conversion of, for example, reflected light from an object using a photoelectric conversion element. The photoelectric conversion element described above has a problem that when outputting an image, the quality of the image deteriorates in a linear or two-dimensional area.

In order to eliminate such problems, the applicant of this application,
Locally change the output level S of a local photoelectric conversion element in a group of photoelectric conversion elements using a scale conversion interval value V-A, ·Xn/ (xn+5n) having predetermined characteristics of the human visual system. Locally adaptive scale conversion method that adjusts by
No. 1748) was proposed. The image signal is output externally through the signal processing system of the conversion circuit (scale conversion circuit) to a brightness value, but due to local illumination unevenness or changes in the intensity of illumination conditions, some parts of the image signal may be insufficiently sensitive or [Problem to be Solved by the Invention] However, in the above-mentioned conventional local adaptation scale conversion method, when n・0.6 is set, the output characteristics after level conversion are similar to the conversion of human vision. This approximates the characteristics, but if the scale conversion equation has an n-th power (n≠1) term, there is a problem in that the scale conversion circuit must be manufactured with high precision. In addition, although this method was able to solve local problems, it was not possible to obtain correct values under illumination conditions where the illumination was uneven (for example, it was bright on the right and darkened toward the left).

SUMMARY OF THE INVENTION An object of the present invention is to improve the conversion characteristics of a scale conversion circuit and to provide a human-powered image device that can obtain accurate brightness values even if the illumination conditions are uneven. Signal output means for outputting the brightness value of the imaging means corresponding to the reference position of the image, the position information of the contour portion obtained from the contour extraction means, and the difference value obtained from the calculation means to the outside as compressed image information. It is characterized by: In addition to the first embodiment, the second embodiment of the present invention provides that the imaging means includes a sensor in which photoelectric conversion elements are arranged for each pixel, and that the output intensity value of each photoelectric conversion element is expressed by the following equation (1). (2)
, [Means for solving the intelligence problem] Based on the outputs of the extraction means, the imaging means, and the contour extraction means,
A calculating means for calculating the difference value between the brightness values of two ground color parts of a subject image with the outline part as a boundary line for each ground color part divided by the outline part, and an average result of averaging the subject brightness values ■. (a variable determined to have a predetermined target value).

[Function] The present invention provides that the difference in brightness between neighboring areas of the ground image is small;
In addition, we focused on the fact that an outline representing the characteristics of the image occurs at the boundary between two background color parts with different brightness, and based on the position information of the outline extracted by the outline extraction means and the outline calculated by the calculation means. The difference value between the brightness values of the two ground color parts sandwiching the area and the brightness value of a specific reference point are output as image information. The brightness value of the pixels up to the position of the outline part is set to the same value as the initial value above, and the pixel positions of other ground color parts beyond the outline part are set by adding the difference value to the above setting value. By setting this value to the brightness value of all pixels on the screen, in other words, image information can be constructed. Even if there is uneven illuminance due to changes in the brightness value, or the similarity of the output characteristics of the signal processing system partially deteriorates, it is possible to construct image information. The ground image part is corrected and a high quality image can be obtained. Furthermore, since the information outputted as image information from the image input device is only the position information of the contour part, the brightness difference value corresponding to the brightness information, and the brightness value of the reference point, the brightness values of all pixels on one screen are There is no need to output the image information, and the amount of image information can be significantly compressed.

Furthermore, in the second embodiment of the present invention, in addition to the first embodiment, the conversion to a brightness value is performed to a brightness that accurately matches the characteristics of human vision, and the sensitivity to changes in brightness is automatically adjusted. [Embodiments] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. Figure 1 shows the functional configuration of an embodiment of the present invention. In FIG. 1, reference numeral 100 denotes an imaging means for reading the brightness value of each pixel of a subject image. Reference numeral 200 denotes a contour extracting means for extracting the position of the contour portion of the subject image based on the brightness value obtained from the imaging means. 300 calculates, based on the outputs of the imaging means and the contour extracting means, the difference value between the brightness values of two ground color parts of the subject image with the contour part as a boundary line for each outer ground color part separated by the contour part. It is a calculation means that calculates.

400 externally outputs the brightness value of the imaging means corresponding to the reference position of the subject image, the position information of the outline part obtained from the outline extraction means, and the difference value obtained from the calculation means as compressed image information. This is a signal output means for outputting to.

Ru. Figure 2 shows a specific circuit configuration of an embodiment of the present invention. In FIG. 2, IO is an image detector that photoelectrically converts light reflected from an object into an image signal of a luminance level, and a plurality of photoelectric conversion elements, such as a solid-state image sensor, which correspond to pixels, are connected in a two-dimensional (planar) manner. It is arranged in S (x) ax, (1+5 .aX/ (x+1-
exe)) m formula (2) (where Am is a constant,
x. is a standard value and is a variable determined so that the average result of averaging the brightness values V for each pixel becomes a predetermined target value). .

30 is a ^/D converter that converts the output signal from the amplifier 20 into digital data. 40 performs scale conversion processing from luminance level to brightness level and illumination unevenness (
Microcomputer 40 that performs shading) correction
It is. The microcomputer 40 is a central processing unit (CPI).
I) 41, read-only memory (ROM), random access memory (RAM) 43 and human output interface (I/O) 44-1. 4-2 is connected to a common bus.

The control procedures shown in FIGS. 3(^) to (D) are stored in the ROM 42, and are read and executed by the CPtl 41. R
The AM 43 is provided with a scale conversion data storage area 43-1, a difference data storage area 43-2, and a general calculation data storage area 43-3. Here, the scale conversion data is image data after performing scale conversion, and the difference data is the slip calculation data storage area 43-3.
(step 5100). Next, scale change 1! will be described later for this image No. 4g using CPII41! llj. (Step 5200)
, difference calculation processing (step 5300) and image compression processing (step 5400), the compressed image signal is output from the I/044-2 to an image processing device (not shown). The operation of the circuit shown in FIG. 2 will be explained with reference to FIG. 3(^).

The image signal for one screen of brightness level transmitted from the image detector 10 via the amplifier 20 and the ^/D converter 30 is sent to CPtl4 1 of the microcomputer 40 from the upper left corner to the right as 1, 2, The numbers of pixel positions are determined in order like 3, 4, and so on.

In FIG. 3(B), first, the magnitude of the first (I-1) luminance signal is converted into brightness. At this time, the following equation is used as the brightness function.

■-＾．・x/ (x+S (x))
...(1)S (x)=x. (1+5.6-x
/(x+1.8xs))...(2) Here, Am is a constant number "20", X is the image data of the luminance signal level at the ■th position + XI is the image data after scale conversion for l screen. The average value R8 is the set value, for example, the middle gray lightness “5”
This is a standard value determined so that In this example, Figure 4 (
Set any P0 in the screen shown in ^) as the reference point,
The image data at this point P0, that is, the magnitude of the luminance signal is set to X. Define as the initial value.

Calculating scale conversion data ■ for the first image data from equations (1) and (2), blocks 5211 to
5201). By repeating this feedback process, when a scale conversion value for each pixel with an overall average brightness of approximately "5" is obtained, the image data is transferred from the calculation data storage area 43-3 to the scale conversion data storage area 43-1. is transferred and stored by the CPU 41 (step 5211-
5213).

FIG. 3(C) shows a detailed control procedure for the difference counting process in step 5300 shown in FIG. 3(A).

This difference counting process is a process of extracting the outline of an image, and once the scale conversion data scale conversion process for a certain pixel position is executed, the brightness average Ra of the scale conversion data is then calculated (step S206→5210).

If the difference between this brightness average Ra and the target value "5" is not within the tolerance range, increase or decrease the value of the set value X by a minute value Δx1 in the direction of approaching the target value, and then The second scale conversion process is repeated from the image data (luminance signal) of and compare each difference value with the threshold value of 0.1.
When the absolute value of this difference value is less than 0.1, the difference value as image data is set to "0", and when it is more than 0.1, it is set as is and is stored in the corresponding storage area of the difference data storage area 43-2. Store (steps 5304 and 5305), for example, image A with brightness 4 and brightness 5 as shown in FIG.
When the above difference value calculation is performed for image B of , image C of brightness 6, and image D of brightness 3, the difference data storage area 43-
When the stored data in 2 is expanded flatly, it becomes as shown in Figure 4 (B). For example, the difference values of the ground image data in image A indicated by line segments Ql-Q3 and Q3-Q8 are all "0".
”, and the border line with other images indicates a difference value other than “0”. Figure 3 (D) shows step 5400 in Figure 3 (^)
The detailed steps of image compression processing are shown below.

First, the reference position PO brightness data is read from the scale conversion storage area 43-1, and this brightness data (in this example, brightness "6") and its pixel position are calculated in the data storage area 43-1.
(Step 5402). Next, the brightness data is read in a certain direction in the difference data storage area 43-2 from the reference point po shown in FIG. Store the information in the calculation data storage area 43-3 as compressed data (step S404→5405). When transferring image information to the image processing device, along with the brightness value of the reference point, the compressed image data closest to the pixel position is sequentially read out from the calculation data storage area 43-3 I/044-
Transfer via 2. On the other hand, on the image processing device side, based on the compressed data, the image data is expanded and stored as normal image data for one screen. For example, to introduce an example of development, the brightness data of the reference point po is stored in the pixel position of PQ in the memory. Next, it is determined whether or not there is position information indicating the pixel position on the right side in the compressed data, and if the determination is positive, the result of adding the difference value to the brightness value of the reference point PO is added to the right side of the reference point. Store it as the adjacent brightness value. Thereafter, one screen's worth of images can be expanded and expanded by sequentially determining the adjacent brightness value based on the brightness value at the current position. An example of performing the various processes explained above is shown below. When an image is input to a band-shaped (line-shaped) colored paper with partially different reflectance under uniform illumination conditions, the reflectance Xi and luminance characteristic Yl as shown in Fig. 5 (^) are obtained. A proportional value is obtained, and the brightness characteristic Zl after scale conversion is also shown in Figure 5 (B
), the brightness is very close to the brightness calculated from the reflectance characteristics in Figure 5 (^). As shown in Figure 5 (C), if the illumination is uneven,
The brightness value will no longer indicate reflectance correctly. In addition, when the signal processing of this embodiment is performed, the scale conversion value includes the slope of the change in illumination, but when the brightness of one image is constructed by further calculating the difference, the value shown in FIG. 5(D) is obtained. Almost the same brightness value as when the illumination conditions are uniform can be obtained.

As explained above, the present invention focuses on the fact that the brightness data of adjacent pixels does not change significantly in the (blank) areas other than the image outline, and uses only the outline of the image in one screen. Extract the brightness difference data. The brightness of the plain color delimited by the outline is a constant value, and in this example, the image A in Figure 4 has a brightness of "4".
By setting image B to have a brightness of "6", even if there is a partial difference in the unevenness in illumination or the signal characteristics of the signal processing system, it is possible to obtain a more effective effect than the conventional method of smoothing out the unevenness in illumination. Furthermore, the image data obtained after such processing can be developed into brightness data for one screen if only the difference brightness data of the contour part, its position data, and the brightness data of the reference position are required. Since the number of differential brightness data, etc. mentioned above is very small compared to the data for each pixel on one screen, it is clear that the transfer time will be shorter than before when input information is transmitted from the image input device to the image processing device. Up to now, we have mainly described the processing after charge conversion using a computer, but it is also possible to perform nonlinear conversion and contour information extraction using an electric circuit.In this case, each photodetector output is 》Attach a scale conversion circuit as shown in
It is desirable to perform parallel processing. This method is described below. Conversion to the brightness scale is performed using equations (1) and (2
), but in order to save on the division circuit, we modified it as follows. In addition, the ratio of the input X to the standard value xe was taken as Z, and the subsequent operation of the circuit was standardized so that the value of 2 normally operated in the range of "0" to 5. The formula is . Here, compare the second term of s(X) and the form of ■ and approximate it using the following formula. However, a is a constant. From these equations, the scale conversion circuit shown in FIG. 6(B) is constructed as an example. Here sa. nY is a constant value determined so that the circuit most closely approximates the brightness conversion function, and is set to 8.2, ny = 0.025. With this circuit, the error in conversion to brightness was less than 0.2, and almost satisfactory results were obtained. In order to obtain accurate brightness for the value of x0, negative feedback may be performed using the method shown in the figure so that the average of the values after conversion becomes the intermediate value of brightness 5, as in the method using the computer described above. However, for convenience, it may take a value proportional to the average value of human power light. The contour information detection circuit is realized by subtracting the outputs of adjacent converted value outputs and using an arithmetic unit that has a dead zone for small signals below the threshold. The processing up to this point is performed in parallel for each photodetector. After this, processing similar to that of the signal output means (400 in Figure 1) is performed. Possible applications of this example are as follows. (1) In this embodiment, the microcomputer 40 is placed at the rear end of the signal processing system to correct uneven illumination and partial deviations in output characteristics. If it is desired to correct non-uniformity in the output characteristics of the conversion element, an analog circuit having the same function as the present invention may be constructed in place of the microcomputer 40, and provided between the amplifier 20 and the image detector lO. ．． At this time, the above-mentioned correction circuit is connected to the image detector 10 and the increaser 111.
20. If an integrated circuit is fabricated as a layered structure together with a scale conversion circuit in place of the microcomputer 40, it can be used, for example, in robot vision or a video camera, and can provide high-quality images. (2) In this embodiment, the formula proposed by the applicant is used as the scale conversion function, but if the luminance level is simply to be nonlinearly transformed, a logarithmic transformation formula or a power transformation formula may be used. You may also use an approximation formula for Munsell's brightness function. (3) In this embodiment, image processing is performed after converting the brightness detected by the image detector into brightness, but the present invention can also be applied to a device that outputs an image signal at a brightness level. In this case, it is preferable to read the brightness data as brightness data and execute the control procedures shown in FIGS. 3(^) to (D) of this embodiment. In this case, the scale conversion process shown in FIG. 3(B) corresponds to the signal level conversion process of the amplifier. [Effects of the Invention] As explained above, in the present invention, the position information of the contour portion extracted by the contour extraction means and the difference value between the brightness values of two ground color portions sandwiching the contour portion, calculated by the calculation means. , and the brightness value of a specific reference point are output as image information, and the image processing device that receives this information uses the brightness value of the reference point position as an initial value and calculates the pixels up to the outline position. By setting the brightness value to the same value as the initial value above, and setting the pixel position of the ground color part beyond the outline part to the value obtained by adding the difference value to the above setting value, the brightness value of all pixels on the screen, i.e. , it is possible to construct image information. As a result, the brightness value of each pixel in the ground color part of the constructed image information becomes a constant value, and there may be irregularities in illuminance due to changes in lighting conditions or similarities in the output characteristics of the signal processing system. Even if it deteriorates, the ground image part will be corrected,
Good quality images can be obtained. Furthermore, since the information outputted as image information from the image input device is only the position information of the contour part, the brightness difference value corresponding to the brightness information, and the brightness value of the reference point, the brightness values of all pixels on one screen are There is no need to output the image information, and the amount of image information can be significantly compressed. Furthermore, in the second embodiment of the present invention, in addition to the first embodiment, a scale conversion formula based on a local adaptation method is used as an accurate brightness conversion formula that adapts to brightness changes, so that the brightness of the entire image of the subject is suitable for signal processing. It is possible to relatively adjust the brightness level, and since the above conversion formula is a linear formula, it is easy to calculate and can be realized with an electric circuit, and high-quality images can be obtained. An explanatory diagram showing the correspondence. Figures 5 (^) to (0) are explanatory diagrams showing the scale conversion characteristics of the embodiment of the present invention. Figure 6 (^) is a block diagram showing another circuit configuration of the embodiment of the present invention. , Figure 6(B) shows the scale conversion circuit 5 shown in Figure 6(^).
Figure 7 is a block diagram showing the functional configuration of a conventional example.

[Brief explanation of drawings]

Fig. 1 is a block diagram showing a functional configuration of an embodiment of the present invention, Fig. 2 is a block diagram showing a specific circuit structure of an embodiment of the present invention, and Figs. 3 (^) to (D) are embodiments of the present invention. Flowchart showing the control procedure executed by the CPII41, Fig. 4 (^)
and (B) is the difference between the subject image of the embodiment of the present invention and the data stored in the differential data storage area 43-2.
Figure 3 (A) Flow chart of the present invention (C) Figure 3 (D) Figure 5 (I & Example) Figure 6 (A) Block diagram showing the principle of the import information detector 7 Figure 6 (A) Figure 7 showing the structure of Kai-mae 1 by Atsushi Ishiba No. 2 Name of the invention image input device 3 Relationship with the person making the amendment Patent applicant 1-114 Kasumigaseki 1-3-1, Chiyoda-ku, Tokyo Director of the Agency of Industrial Science and Technology Ken Sugiura 4 Designated agent 7 Contents of the amendment 1) In the specification , page 4, line 9, "r62-11748" is corrected to r62-117478J.

Claims (1)

[Scope of Claims] 1) Imaging means for reading the brightness value of each pixel of the object image; and contour extraction means for extracting the position of the outline part of the object image based on the brightness value obtained from the image sensing means. , Based on the outputs of the imaging means and the contour extracting means, calculate the difference value between the brightness values of two ground color parts of the subject image with the outline part as a boundary line for each ground color part divided by the outline part. a calculation means for calculating a brightness value of the imaging means corresponding to a reference position of the subject image, positional information of the contour portion obtained from the contour extraction means, and the difference value obtained from the calculation means. An image input device comprising: signal output means for outputting to the outside as compressed image information. 2) In the image input device according to claim 1, the image capturing means includes a sensor in which photoelectric conversion elements are arranged for each pixel, and when the output intensity value of each of the photoelectric conversion elements is X, the following equation (1) is satisfied.
, (2), V=A_m・x/(x+S(x))...Formula (1) S(
x)=x_(1+5.6x/(x+1.8x_s))・
...Equation (2) (where A_m is a constant, x_s is a standard value, and is a variable determined so that the result of averaging the brightness values V for each pixel becomes a predetermined target value). An image input device characterized by having an output section that outputs an output.