JP2019110443A - Image processing apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processing device capable of detecting a false color with a simple structure and correcting the color so as to reduce the false color. An overexposed pixel determination unit 21 determines whether each pixel in image data is an overexposed pixel. The overexposure pixel count accumulating unit 22 accumulates the number of pixels determined to be overexposure pixels. The overexposure pixel number threshold determination unit 23 determines whether the cumulative overexposure pixel number exceeds the overexposure pixel number threshold. The B threshold determination unit 24 determines whether or not the B signal is larger than the B threshold. The (BG) threshold determination unit 25 determines whether or not the (BG) value is larger than the (BG) threshold. The correction necessity determination unit 26 determines that the cumulative number of overexposed pixels exceeds the overexposed pixel threshold, determines that the B signal is greater than the B threshold, and the (BG) value is greater than the (BG) threshold. If it is determined that the color of the pixel of interest needs to be corrected. The color correction unit corrects the color signal when it is determined that the color of the pixel of interest needs to be corrected. [Selection diagram] Fig. 3

Description

  The present invention relates to an image processing apparatus that detects and reduces false colors generated in a captured image.

  When an object including overexposure such as back light is imaged by an imaging device equipped with a high-magnification optical system using a low-cost lens, the low-intensity part is compared to the overexposure part adjacent to the overexposure part. Saturated purple or blue false colors may occur. This type of false color is called purple fringe. Patent Document 1 describes an image processing apparatus that detects and reduces purple fringes.

Unexamined-Japanese-Patent No. 2015-211325

  There is a need for an image processing apparatus capable of detecting false colors and correcting colors with a simple configuration. An object of the present invention is to provide an image processing apparatus capable of detecting a false color with a simple configuration and capable of correcting the color so as to reduce the false color.

  According to the present invention, an overexposure pixel determination unit that determines whether or not an overexposure pixel is present for each pixel in image data configured by the input R, G, and B signals; Alternatively, the overexposure pixel count accumulation section accumulates the number of pixels determined to be the overexposure pixels by the overexposure pixel determination section in the partial area of the frame, and the overexposure pixel count accumulation section An overexposure pixel number threshold determination unit that determines whether the accumulated overexposure pixel count exceeds the overexposure pixel count threshold, a B threshold determination unit that determines whether the B signal in the pixel of interest is greater than the B threshold, and (B−G) a threshold determination unit that determines whether a (B−G) value obtained by subtracting a G signal from a B signal in the target pixel is greater than a (B−G) threshold, and the white overexposure pixel number threshold Cumulative whiteout pixel count is white in the judgment unit And the B threshold determination unit determines that the B signal at the target pixel is greater than the B threshold, and the (BG) threshold determination unit determines the (BG) value at the target pixel. Is determined to be larger than the (B−G) threshold value, it is determined that the color of the pixel of interest is required to be corrected assuming that false color is generated in the pixel of interest, and a correction necessity determination unit; A color correction unit that corrects at least the G signal and the B signal in the target pixel so as to reduce the false color occurring in the target pixel when the color of the target pixel is determined to be corrected in the determination unit; To provide an image processing apparatus characterized by comprising:

  According to the image processing apparatus of the present invention, purple fringes can be detected with a simple configuration, and colors can be corrected.

FIG. 1 is a block diagram showing an image processing apparatus according to a first embodiment. It is a flowchart which shows operation | movement of the image processing apparatus of 1st Embodiment. It is a block diagram showing the image processing device of a 2nd embodiment. It is a flowchart which shows operation | movement of the image processing apparatus of 2nd Embodiment.

  Hereinafter, the image processing apparatus according to the first and second embodiments will be described with reference to the attached drawings.

First Embodiment
Purple fringing, which is a false color of purple or blue with high saturation, is a G signal around the pixel where overexposure occurs among the color signals of R (red), G (green), and B (blue). Is less than a certain value, and it can be said that the B signal is larger than the G signal by a certain amount or more. The image processing apparatus 100 according to the first embodiment determines that a state in which the G signal is less than a certain amount and the B signal is larger than the G signal by a certain amount or more is purple fringe, and corrects color to reduce purple fringes.

  In the image processing apparatus 100 of the first embodiment shown in FIG. 1, a captured image generated by capturing an image of a subject by the imaging device is supplied as input image data. The input image data is a digital image signal of a still image or a moving image constituted by R, G and B signals. The input image data is quantized by 8 bits, and each pixel value takes any value from 0 to 255. The pixel values of the R, G, and B signals of the pixel of interest are referred to as an input R value, an input G value, and an input B value, respectively.

  In FIG. 1, R, G, and B signals are input to the peripheral overexposure pixel count unit 11. In the block of (N × N) pixels including the pixel of interest and the peripheral pixels centered on the pixel of interest, the peripheral white skipping pixel count unit 11 determines all of the input R value, the input G value, and the input B value. The pixel exceeding the threshold value is determined to be a pixel in the overexposure state, and the number of pixels is counted. The threshold is, for example, 220. N may be selected from 3 to 100 depending on the computational resources or the image size.

  The count value wh_num by the peripheral overexposure pixel count unit 11 is supplied to the G threshold value determination unit 12 and the (B−G) threshold value determination unit 13. The G threshold determination unit 12 determines a G threshold, which is a threshold to be compared with the input G value, based on A × wh_num + B.

  As an example, when N is 51, any number of 0.03 to 0.1 may be selected as A, and any number of 50 to 150 may be selected as B. The reason why A is a positive number is to increase the G threshold value according to the count value wh_num because the possibility that the input G value increases as the number of overexposed pixels increases around the target pixel is high.

  (B−G) The threshold value determination unit 13 determines a (B−G) threshold value that is a threshold value to be compared with a (B−G) value obtained by subtracting the input G value from the input B value based on C × wh_num + D. .

  As an example, when N is 51, C may be any number from -0.1 to -0.02, and D may be any number from 60 to 120. The reason why C is a negative number is that as the number of whitening pixels increases around the pixel of interest, it is likely that the input G value and the input B value approach saturation, and the (B−G) value decreases. This is to reduce the (B−G) threshold according to the count value wh_num.

  The G threshold determination unit 14 is supplied with the G signal and the G threshold. The (B−G) threshold determination unit 15 is supplied with the G signal, the B signal, and the (B−G) threshold. The G threshold value determination unit 14 determines whether or not the input G value <G threshold value. (BG) The threshold determination unit 15 determines whether or not (BG) value> (BG) threshold. The determination results by the G threshold determination unit 14 and the (B−G) threshold determination unit 15 are supplied to the correction necessity determination unit 16.

  The correction necessity determination unit 16 determines that the input G value <G threshold in the G threshold determination unit 14, and the (B−G) value> (B−G) threshold in the (B−G) threshold determination unit 15. When it is determined that there is purple fringes in the target pixel, it is determined that the target pixel is required to be corrected. The determination result of necessity of correction by the correction necessity determination unit 16 is supplied to the B correction unit 17 and the G correction unit 18 which are color correction units. A G signal and a B signal are supplied to the B correction unit 17 and the G correction unit 18.

  When the correction necessity determination unit 16 determines that the correction is necessary, the B correction unit 17 outputs an addition value obtained by adding the k × (B−G) threshold to the input G value as a correction B value. The correction B value is a first addition value obtained by adding a first multiplication value obtained by multiplying the (G) threshold value by k, which is a first constant, to the input G value. The correction B value is output from the image processing apparatus 100 and is also supplied to the G correction unit 18. The constant k may be determined in consideration of the overall color balance after correction as a value of about 0.3 to 0.9.

  When the correction necessity determination unit 16 determines that correction is unnecessary, the B correction unit 17 outputs the input B value as it is.

  By the correction of the B value by the B correction unit 17, it is possible to obtain a B value in which the difference between the corrected B value and the input G value does not exceed the (B−G) threshold. When the coefficient k is 1.0, the difference between the modified B value and the input G value matches the (B−G) threshold.

  When the correction necessity determination unit 16 determines that correction is necessary, the G correction unit 18 multiplies the input G value by the difference between the input B value and the correction B value by the constant p, p {input B value An addition value obtained by adding − (input G value + k × (B−G) threshold value)} is output as a corrected G value. The correction G value is a second addition value obtained by adding a second multiplication value obtained by multiplying the difference between the input B value and the correction B value by the second constant to the input G value. The corrected G value is output from the image processing apparatus 100. The constant p may be determined in consideration of the overall color balance after correction as a value of about 0.05 to 0.3.

  When the correction necessity determination unit 16 determines that the correction is unnecessary, the G correction unit 18 outputs the input G value as it is.

  Due to the correction of the G value by the G correction unit 18, the target pixel with the input B value close to saturation and the small input G value can be prevented from being excessively strongly corrected for redness, and it may be closer to daylight white it can.

  In FIG. 1, the input R value is output from the image processing apparatus 100 as it is. When purple fringes occur, the input G value and the input B value are corrected to output a corrected G value and a corrected B value. The input R value, the corrected G value, and the corrected B value are output from the image processing apparatus 100 as output image data. If purple fringes do not occur, the input R value, the input G value, and the input B value are output from the image processing apparatus 100 as output image data as they are.

  Although not shown in FIG. 1, the input R value, the corrected G value and the corrected B value (or the input R value, the input G value and the input B value) are mutually matched in timing as required. It is output. Either the input R value is delayed by the time required to generate the corrected G value and the corrected B value, or the input R value, the corrected G value, and the corrected B value are temporarily stored in the storage unit and output.

  The operation (image processing method) of the image processing apparatus 100 according to the first embodiment will be further described using the flowchart shown in FIG. In FIG. 2, when the image processing apparatus 100 starts operation, image data is taken in at step S100. The image data may be a still image or a moving image.

  In step S101, the peripheral overexposure pixel count unit 11 counts the number of overexposure pixels around the pixel of interest. In step S102, the G threshold value determination unit 12 determines the G threshold value according to the number of overexposed pixels. (B-G) The threshold value determination unit 13 determines the (B-G) threshold value in accordance with the number of overexposure pixels in step S103. The order of step S102 and step S103 may be reversed or may be simultaneous.

  The G threshold value determination unit 14 determines whether or not the input G value <G threshold value in step S104. (B−G) The threshold value determination unit 15 determines whether or not (B−G) value> (B−G) threshold in step S105. The order of step S104 and step S105 may be reversed or may be simultaneous.

  In step S106, the correction necessity determination unit 16 determines whether or not the pixel of interest needs correction. The correction necessity determination unit 16 determines that the target pixel is required to be corrected when the input G value is less than the G threshold and the (BG) value is greater than the (BG) threshold. If this condition is not satisfied, the correction necessity determination unit 16 determines that the target pixel is not required to be corrected.

  If the pixel of interest is required to be corrected (YES), the B correction unit 17 outputs an addition value obtained by adding the k × (B−G) threshold to the input G value as a correction B value in step S107. In step S108, the G correction unit 18 outputs, as a corrected G value, an added value obtained by adding p {input B value − (input G value + k × (B−G) threshold)} to the input G value. The processing device 100 shifts the processing to step S119. The order of step S107 and step S108 may be reversed or may be simultaneous.

  On the other hand, if the pixel of interest is not required to be corrected at step S106 (NO), the B correction unit 17 outputs the input B value as it is at step S117, and the G correction unit 18 receives the input G at step S118. The value is output as it is, and the image processing apparatus 100 shifts the processing to step S119. The order of step S117 and step S118 may be reversed or may be simultaneous.

  The image processing apparatus 100 determines in step S119 whether or not the pixel of interest is the final pixel, and if it is not the final pixel (NO), repeats the processing after step S100, and if it is the final pixel (YES), End the process. The final pixel is the last pixel in the frame if it is a still image, and the last pixel in the last frame when the input image data is stopped if it is a moving image.

  In the image processing apparatus 100 according to the first embodiment described above, purple fringes can be accurately detected by a simple configuration that detects the presence or absence of purple fringes by the peripheral overexposure pixel count unit 11 to the correction necessity determination unit 16. Can.

  Instead of the B correction unit 17 and the G correction unit 18 correcting the input B value and the input G value as described above, the input B value and the input G value are corrected by another correction method such as hole filling processing or saturation reduction processing. You may When the B correction unit 17 and the G correction unit 18 correct the input B value and the input G value as described above, the purple fringe can be corrected with almost no deterioration in the image quality. Therefore, the B correction unit 17 and the G correction unit 18 It is preferable to correct the input B value and the input G value as follows.

Second Embodiment
Purple fringing means that the saturation level of the B signal is high, so if the B signal is above a certain level and the B signal is above the G signal by a certain level or more around the pixel where overexposure occurs. It can also be said. The image processing apparatus 200 according to the second embodiment determines that a state in which the B signal is at least a predetermined level and the B signal is at least a predetermined level larger than the G signal is purple fringe, and corrects color to reduce purple fringe.

  In the first embodiment, the number of white-out pixels in the peripheral pixels of the pixel of interest is counted, but in the second embodiment, the number of pixels exceeding a certain number of pixels in one frame is white. In the case of a pixel in a state of skipping, it is assumed that overexposure has occurred.

  In the image processing apparatus 200 according to the second embodiment shown in FIG. 3, a captured image generated by the imaging apparatus capturing an image of a subject is supplied as input image data. In the image processing apparatus 200 of the second embodiment, the input image data is a digital image signal of a moving image constituted by R, G and B signals. The input image data is quantized by 8 bits, and each pixel value takes any value from 0 to 255. The pixel values of the R, G, and B signals of the pixel of interest are referred to as an input R value, an input G value, and an input B value, respectively.

  In FIG. 3, the overexposure pixel determination unit 21 determines whether all of the input R value, the input G value, and the input B value exceed a predetermined threshold for each pixel, and whites all pixels exceeding the threshold. It is determined that the pixel is in the jump state. The threshold is, for example, 220. The determination result of the pixel unit by the overexposure pixel determination unit 21 is supplied to the overexposure pixel number accumulation unit 22. The overexposure pixel count accumulation unit 22 accumulates the number of pixels determined to be the pixels in the overexposure state in each frame. The accumulated overexposure pixel count by the overexposure pixel count accumulation unit 22 is supplied to the overexposure pixel count threshold determination unit 23.

  The overexposure pixel number accumulation unit 22 may accumulate the number of pixels determined to be the pixels in the overexposure state in the entire frame, or it is determined to be the pixels in the overexposure state within a partial area of the frame The number of pixels may be accumulated.

  The overexposure pixel number threshold determination unit 23 determines whether the accumulated overexposure pixel number exceeds the overexposure pixel number threshold. The overexposure pixel number threshold may be set to about several percent (1 to 9%) of the total number of pixels of the frame. The determination result by the overexposure pixel number threshold determination unit 23 is supplied to the correction necessity determination unit 26.

  Since the overexposure pixel count accumulation unit 22 obtains the accumulated overexposure pixel count in frame units, the timing when the overexposure pixel count threshold determination unit 23 supplies the determination result by the overexposure necessity determination unit 26 is one frame. After the pixel is input. Therefore, it is the next frame that the correction necessity determination unit 26 determines whether to correct the pixel of interest using the determination result by the overexposure pixel number threshold determination unit 23. That is, the correction necessity determination unit 26 determines whether to correct the pixel of interest using the determination result in the previous frame.

  The B threshold value determination unit 24 is supplied with the B signal. In the B threshold determination unit 24, a B threshold which is a threshold to be compared with the input B value is set. The B threshold may be, for example, about 120. The B threshold determination unit 24 determines whether or not input B value> B threshold. The B threshold value determination unit 24 determines whether or not the input B value> B threshold value in the current frame corresponding to the above-mentioned next frame. The determination result by the B threshold determination unit 24 is supplied to the correction necessity determination unit 26.

  (B−G) The threshold value determination unit 25 is supplied with the G signal and the B signal. In the (B−G) threshold value determination unit 25, a (B−G) threshold value is set which is a threshold value to be compared with the (B−G) value. The threshold value (BG) may be, for example, about 80. (B−G) The threshold determination unit 25 determines whether (B−G) value> (B−G) threshold. (B-G) The threshold determination unit 25 determines whether or not (B-G) value> (B-G) threshold in the current frame. (B−G) The determination result by the threshold value determination unit 25 is supplied to the correction necessity determination unit 26.

  The correction necessity determination unit 26 determines that the overexposure pixel count threshold determination unit 23 determines that the total overexposure pixel count> overexposure pixel count threshold and the B threshold determination unit 24 determines that the input B value> the B threshold, (B-G) When it is determined by the threshold value determination unit 25 that (B-G) value> (B-G) threshold value, purple fringe is generated in the target pixel and the target pixel is determined to be corrected. The determination result of the necessity of correction by the correction necessity determination unit 26 is supplied to the B correction unit 27, the G correction unit 28, and the R correction unit 29, which are color correction units.

  A G signal and a B signal are supplied to the B correction unit 27 and the G correction unit 28. The R correction unit 29 is supplied with an R signal.

  When the correction necessity determination unit 26 determines that the correction is necessary, the B correction unit 27 outputs an addition value obtained by adding the k × (B−G) threshold to the input G value as a correction B value. The correction B value is a first addition value obtained by adding a first multiplication value obtained by multiplying the (G) threshold value by k, which is a first constant, to the input G value. The correction B value is output from the image processing apparatus 200 and is also supplied to the G correction unit 28 and the R correction unit 29. The constant k may be determined in consideration of the overall color balance after correction as a value of about 0.3 to 0.9.

  When the correction necessity determination unit 26 determines that the correction is unnecessary, the B correction unit 27 outputs the input B value as it is.

  By the correction of the B value by the B correction unit 27, it is possible to obtain a B value in which the difference between the corrected B value and the input G value does not exceed the (B−G) threshold. When the coefficient k is 1.0, the difference between the modified B value and the input G value matches the (B−G) threshold.

  When it is determined that correction is necessary in the correction necessity determination unit 26, the G correction unit 28 multiplies the difference between the input B value and the correction B value by the constant p to the input G value, p {input B value An addition value obtained by adding − (input G value + k × (B−G) threshold value)} is output as a corrected G value. The correction G value is a second addition value obtained by adding a second multiplication value obtained by multiplying the difference between the input B value and the correction B value by the second constant to the input G value. The corrected G value is output from the image processing apparatus 200. The constant p may be determined in consideration of the overall color balance after correction as a value of about 0.05 to 0.3.

  When the correction necessity determination unit 26 determines that the correction is unnecessary, the G correction unit 28 outputs the input G value as it is.

  Due to the correction of the G value by the G correction unit 28, the target pixel with the input B value close to saturation and the small input G value can be prevented from being excessively strongly corrected for redness, and it may be closer to daylight white it can.

  When it is determined that correction is necessary in the correction necessity determination unit 26, the R correction unit 29 multiplies the difference between the input B value and the correction B value from the input R value by the constant q, q {input B value The subtraction value obtained by subtracting − (input G value + k × (B−G) threshold value)} is output as a corrected R value. The correction R value is a subtraction value obtained by subtracting the third multiplication value obtained by multiplying the difference between the input B value and the correction B value by the third constant from the input R value. The corrected R value is output from the image processing apparatus 200. The constant q may be determined in consideration of the overall color balance after correction as a value of about 0.05 to 0.1.

  When the correction necessity determination unit 26 determines that the correction is unnecessary, the R correction unit 29 outputs the input R value as it is.

  The R value is corrected by the R correction unit 29 such that the input B value is close to saturation, and the reddish intensity still remains even if the target pixel with a small input G value is corrected to be close to daylight white. Since it reduces, it becomes possible to correct to a more natural color.

  In FIG. 3, when purple fringes occur, the corrected R value, corrected G value and corrected B value are output from the image processing apparatus 200 as output image data, and if purple fringes do not occur, the input R value, The input G value and the input B value are output from the image processing apparatus 200 as output image data as they are.

  Similarly, in FIG. 3, the corrected R value, the corrected G value, and the corrected B value (or the input R value, the input G value and the input B value) are output in timing with each other as needed.

  The operation (image processing method) of the image processing apparatus 200 of the second embodiment will be further described using the flowchart shown in FIG. In FIG. 4, when the image processing apparatus 200 starts operation, image data is fetched in step S200. The image data here is a moving image.

  In step S201, the overexposure pixel determination unit 21 determines whether each pixel is an overexposure pixel. In step S202, the overexposure pixel count accumulation unit 22 accumulates the overexposure pixel count in units of frames. In step S203, the overexposure pixel count threshold determination unit 23 determines whether the total overexposure pixel count> overexposure pixel count threshold is satisfied.

  The B threshold value determination unit 24 determines whether or not the input B value> B threshold value in step S204. (B−G) The threshold value determination unit 25 determines whether or not (B−G) value> (B−G) threshold in step S205.

  In step S206, the correction necessity determination unit 26 determines whether or not the pixel of interest needs correction. The correction necessity determination unit 26 determines the pixel of interest when the cumulative overexposure pixel number> overexposure pixel number threshold value, and the input B value> B threshold value and the (B−G) value> (B−G) threshold value. Is determined to be necessary.

  If the pixel of interest is required to be corrected (YES), the B correction unit 27 outputs an addition value obtained by adding the k × (B−G) threshold to the input G value as a correction B value in step S207. In step S208, the G correction unit 28 outputs an added value obtained by adding p {input B value− (input G value + k × (B−G) threshold value)} to the input G value as a corrected G value. In step S209, the R correction unit 29 outputs a subtraction value obtained by subtracting q {input B value − (input G value + k × (BG) threshold)} from the input R value as a corrected R value, The processing device 200 shifts the processing to step S210.

  The order of step S207 to step S209 is arbitrary and may be simultaneous.

  On the other hand, if the target pixel is not required to be corrected in step S206 (NO), the B correction unit 27 outputs the input B value as it is in step S217. In step S218, the G correction unit 18 outputs the input G value as it is. In step S219, the R correction unit 29 outputs the input R as it is, and the image processing apparatus 200 shifts the processing to step S210.

  The order of step S217 to step S219 is arbitrary and may be simultaneous.

  The image processing apparatus 200 determines in step S210 whether or not the pixel of interest is the final pixel, and if it is not the final pixel (NO), repeats the processing after step S200, and if it is the final pixel (YES), End the process. The last pixel is the last pixel in the last frame when the input image data stops.

  In the image processing apparatus 200 of the second embodiment described above, purple fringes can be accurately detected by a simple configuration that detects the presence or absence of purple fringes by the overexposure pixel determination unit 21 to the correction necessity determination unit 26. .

  Instead of the B correction unit 27, the G correction unit 28, and the R correction unit 29 correcting the input B value, the input G value, and the input R value as described above, other correction methods such as hole filling processing or saturation reduction processing The input B value, the input G value, and the input R value may be corrected by When the B correction unit 27, the G correction unit 28, and the R correction unit 29 correct the input B value, the input G value, and the input R value as described above, the purple fringe can be corrected with almost no image quality deterioration. It is preferable that the correction unit 27, the G correction unit 28, and the R correction unit 29 correct the input B value, the input G value, and the input R value as described above.

<Modification>
In the image processing apparatus 100 according to the first embodiment, it is determined whether overexposure occurs or not based on the number of overexposure pixels in the peripheral pixels of the target pixel. This will be referred to as a first overexposure generation determination method. In the image processing apparatus 200 of the second embodiment, it is determined whether overexposure has occurred based on the number of overexposure pixels in the previous frame. This will be referred to as a second whiteout occurrence determination method.

  In the image processing apparatus 100 according to the first embodiment, when the G signal is less than a certain value and the B signal is larger than the G signal by a certain amount or more, it is determined that purple fringes are generated. This is the first purple fringe determination method. In the image processing apparatus 200 according to the second embodiment, it is determined that purple fringes occur when the B signal is equal to or more than a predetermined level and the B signal is larger than the G signal by a predetermined number or more. This is the second purple fringe determination method.

  In the image processing apparatus 100 of the first embodiment, the G threshold and the (B-G) threshold are varied according to the number of overexposed pixels, and in the image processing apparatus 200 according to the second embodiment, the number of overexposed pixels is The B threshold and the (B-G) threshold are constant regardless of. In the image processing apparatus 100 of the first embodiment, the G and B signals are corrected, and in the image processing apparatus 200 of the second embodiment, the R, G and B signals are corrected.

  The image processing apparatus 100 according to the first embodiment has a configuration in which a combination of a first overexposure determination method, a first purple fringe determination method, threshold variation, and correction of G and B signals is adopted. The image processing apparatus 200 according to the second embodiment has a configuration in which a combination of a second overexposure determination method, a second purple fringe determination method, a constant threshold, and correction of R, G, and B signals is employed.

  The image processing apparatus can adopt any combination other than the combination of the first and second embodiments as a modification.

  The present invention is not limited to the first and second embodiments or modifications described above, and various modifications can be made without departing from the scope of the present invention. The image processing apparatus may be configured by hardware or software, and the use of hardware and software is optional.

  The image processing apparatus may be configured by an integrated circuit (LSI) or may be configured by a central processing unit (CPU) and a memory of a computer. The image processing apparatus 100 of the first embodiment, the image processing apparatus 200 of the second embodiment, or the image processing apparatus of the modification is configured by causing a CPU to execute a computer program (image processing program) stored in a memory. It may be done. The image processing program may be stored and provided on a non-transitory storage medium.

11 peripheral overexposure pixel count unit 12 G threshold determination unit 13 (BG) threshold determination unit 14 G threshold determination unit 15, 25 (BG) threshold determination unit 16, 26 correction necessity determination unit 17, 27 B Correction unit (color correction unit)
18,28 G correction unit (color correction unit)
21 overexposure pixel determination unit 22 overexposure pixel count accumulation unit 23 overexposure pixel count threshold determination unit 24 B threshold determination unit 29 R correction unit (color correction unit)
100, 200 image processing device

G threshold determination unit 1 4, at step S104, determines whether the input G value <G threshold. (B−G) The threshold value determination unit 15 determines whether or not (B−G) value> (B−G) threshold in step S105. The order of step S104 and step S105 may be reversed or may be simultaneous.

Claims (4)

An overexposure pixel determination unit that determines whether each pixel in the image data configured by the input R, G, and B signals is an overexposure pixel;
An overexposure pixel number accumulation unit which accumulates the number of pixels determined as the pixels in the overexposure state by the overexposure pixel determination unit within the frame or a partial region of the frame in the image data;
An overexposure pixel number threshold determination unit that determines whether the cumulative overexposure pixel number obtained in the overexposure pixel number accumulation unit exceeds the overexposure pixel number threshold;
A B threshold determination unit that determines whether or not the B signal at the pixel of interest is greater than the B threshold;
(B−G) threshold determination unit that determines whether (B−G) value obtained by subtracting G signal from B signal in the target pixel is larger than (B−G) threshold;
The overexposure pixel number threshold determination unit determines that the total overexposure pixel number exceeds the overexposure pixel number threshold, and the B threshold determination unit determines that the B signal in the target pixel is greater than the B threshold, G) When it is determined that the (BG) value in the target pixel is larger than the (BG) threshold in the threshold determination unit, the color of the target pixel is determined as false color being generated in the target pixel A correction necessity determination unit that determines that correction is necessary,
When the correction necessity determination unit determines that the color of the target pixel is required to be corrected, at least the G signal and the B signal in the target pixel are corrected so as to reduce the false color occurring in the target pixel. A color correction unit,
An image processing apparatus comprising:   The color correction unit sets a first addition value obtained by adding a first multiplication value obtained by multiplying the (B−G) threshold value by the first constant to the G signal in the target pixel as a corrected B value. The image processing apparatus according to claim 1, characterized in that   The color correction unit adds a second addition value obtained by adding a second multiplication value obtained by multiplying a difference between the B signal in the pixel of interest and the corrected B value to a G signal in the pixel of interest and a second constant. The image processing apparatus according to claim 1, wherein the correction G value is used. The color correction unit corrects an R signal in addition to the G signal and the B signal in the target pixel,
The color correction unit corrects a subtraction value obtained by subtracting a third multiplication value obtained by multiplying a difference between the B signal at the target pixel and the corrected B value from the R signal at the target pixel by a third constant. The image processing apparatus according to any one of claims 1 to 3, characterized in that:

Images