JP2006148748A - Pixel defect correction apparatus and pixel defect correction method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a pixel defect correction device capable of obtaining a high-quality image with little deterioration in image quality by accurately detecting and correcting pixel defects.
A pixel defect correcting apparatus 10 includes an image signal input unit 11 that inputs a video, and the same color pixels around a pixel to be inspected as to whether or not it is a defective pixel in a video input to the image signal input unit 11. Surrounding same color pixel maximum value calculating unit 12 and surrounding same color pixel minimum value calculating unit 13 for calculating a change amount of the value, ambient brightness value calculating unit 14 for calculating a brightness value of pixels around the inspection target pixel, and the same color pixel value A threshold value calculation unit 15 that calculates a threshold value for detecting a defective pixel based on the change amount and the luminance value, a comparator 16 that determines whether the inspection target pixel is a defective pixel using the threshold value, and a comparator And a selector 17 for correcting the pixel value of the defective pixel detected in 16.
[Selection] Figure 1

Description

  The present invention relates to digital image processing such as an imaging apparatus, and more particularly to an apparatus for detecting and correcting pixel defects in a solid-state imaging device.

  It is known that a solid-state imaging device has a defective pixel (scratch) that outputs an output unrelated to the amount of incident light. This appears in the image as black scratches with black spots and white scratches with white spots. In addition, when the temperature of the device rises, abnormal charge accumulation occurs, and scratches called temperature scratches may occur in the image.

  It is required to perform correction processing on defective pixels to suppress deterioration in image quality. There are two methods for correcting defective pixels: a method in which coordinate information of defective pixels is recorded in a nonvolatile memory in advance and a pixel at the coordinates is corrected; and a method in which defective pixels are detected from surrounding image information and corrected.

  Patent Document 1 discloses a method of detecting and correcting defective pixels from surrounding image information. In the apparatus described in this document, when determining whether or not the inspection target pixel is a defective pixel, the frequency of the surrounding pixel having a color filter different from the color of the inspection target pixel is used. Check that the pixel has no high frequency components. Thereafter, when it is detected that the pixel to be inspected has a high-frequency component, it is detected that the pixel is flawed.

  Specifically, band filtering is performed on pixels of a color different from the color of the pixel to be inspected that performs the first process of the first stage. This absolute value is obtained as a second process. This result is compared as a third process to determine whether it is smaller than a threshold set from the outside. In this case, the color signals other than the pixel to be inspected are inspected independently.

  This first stage of inspection confirmed that there was no high frequency component around the pixel to be inspected. In general, it is known that a low-frequency component has a high correlation between other colors, and it can be assumed that the inspection target pixel does not have a high-frequency component.

The second stage inspection to be performed next is to examine the frequency characteristics of the inspection target pixel itself. In addition, when the pixel to be inspected has white scratches, this signal level should be higher than other pixels, so that it is not necessary to take an absolute value as in the first stage. In other words, if all of these inspection items are cleared as a result of subtracting the values of the surrounding pixels from the value of the inspection target pixel and comparing with the externally set threshold value, it is assumed that the inspection target pixel is white.
JP 2001-86517 A (page 3)

  However, as in Patent Document 1, in order to detect a pixel defect from image information around an inspection target pixel, it is necessary to set a determination threshold value in advance. However, it is normal for images with different imaging conditions to have different brightness and sharpness, and there are usually parts with different brightness and sharpness within a single image, all of which have thresholds that allow accurate determination. Is difficult to set.

  For example, when a dark subject is captured, even a pixel defect with a small amount of signal is very noticeable. If the threshold value is lowered in order to detect this, the edge portion of the image is erroneously detected in a bright image, causing image quality deterioration due to erroneous correction.

  The present invention has been made in view of the above background, and is capable of detecting and correcting defective pixels from a dark part to a bright part of an image and from a flat part to an edge part without erroneous detection. The purpose is to provide.

  The pixel defect correction apparatus according to the present invention includes a video input unit that inputs video, and a change amount of the same color pixel value around the pixel to be inspected as to whether the pixel is a defective pixel in the video input to the video input unit. A change amount calculating unit for calculating, a luminance value calculating unit for calculating a luminance value of pixels around the inspection target pixel, and a defective pixel based on the change amount of the same color pixel value and the luminance value; Threshold value calculating means for calculating a threshold value, defective pixel detecting means for determining whether the inspection target pixel is a defective pixel using the threshold value, and correcting a pixel value of the defective pixel detected by the defective pixel detecting means And a correction means.

  As described above, the threshold value used for determining whether or not the inspection target pixel is a defective pixel is calculated based on the change amount of the surrounding same-color pixel value and the luminance value, and thus varies depending on the shooting condition and the like. A threshold value corresponding to the characteristics of the video can be obtained. Accordingly, it is possible to realize a pixel defect correction apparatus that detects and corrects pixel defects without erroneous detection from a dark part to a bright part of an image and from a flat part to an edge part.

  The pixel defect correction apparatus includes recording means for recording coordinates of defective pixels of an image sensor provided in the video input means, and the correction means calculates pixel values of defective pixels at coordinates recorded in the recording means. It may be corrected.

  With this configuration, for example, by recording in advance the coordinates of a defective pixel that has a large effect on the video, it is possible to reliably correct defective pixels that are missed by the threshold value calculated by the threshold value calculation unit.

  In the above-described pixel defect correction device, the threshold value calculation unit may calculate a threshold value larger when the change amount of the same color pixel value is large and larger when the change amount is small.

  In this way, by using a large threshold value for pixels in a portion where the amount of change is large, erroneous detection of the edge portion of the image can be reduced.

  In the pixel defect correction apparatus, the threshold value calculation unit may calculate a threshold value that is larger when the luminance value is small than when the luminance value is small.

  With this configuration, it is possible to reduce the erroneous detection of the edge portion of the image in the case of a bright image with a large luminance value, and to detect the pixel defect with high accuracy in the case of a dark image with a small luminance value.

  According to another aspect of the present invention, there is provided a pixel defect correction apparatus including: a video input unit that inputs a video; and a same color pixel around a pixel to be inspected as a defective pixel in the video input to the video input unit A change amount calculating means for calculating a change amount of the value; a luminance value calculating means for calculating a brightness value of pixels around the inspection target pixel; and the inspection based on the change amount of the same color pixel value and the brightness value. The apparatus includes a defective pixel detection unit that determines whether the target pixel is a defective pixel, and a correction unit that corrects the pixel value of the defective pixel detected by the defective pixel detection unit.

  In this way, by detecting a defective pixel based on the amount of change in the surrounding same-color pixel value and the luminance value, the defective pixel can be detected appropriately according to the characteristics of the video. Accordingly, it is possible to realize a pixel defect correction apparatus that detects and corrects pixel defects without erroneous detection from a dark part to a bright part of an image and from a flat part to an edge part.

  The pixel defect correction method of the present invention includes a video input step for inputting video, and a change amount of the same color pixel value around a pixel to be inspected as to whether the pixel is a defective pixel in the video input in the video input step. A change amount calculating step for calculating, a luminance value calculating step for calculating a luminance value of pixels around the inspection target pixel, and a defective pixel based on the change amount of the same color pixel value and the luminance value A threshold calculating step for calculating a threshold; a defective pixel detecting step for determining whether the inspection target pixel is a defective pixel using the threshold; and a correction for correcting a pixel value of the defective pixel detected in the defective pixel detecting step And a step.

  With this configuration, a pixel defect correction method that detects and corrects pixel defects without false detection from a dark part to a bright part of an image and from a flat part to an edge part as in the pixel defect correction apparatus of the present invention is realized. be able to. Various configurations of the pixel defect correction apparatus of the present invention can be applied to the pixel defect correction method of the present invention.

  According to the present invention, the threshold used for determining whether or not the inspection target pixel is a defective pixel is calculated based on the change amount of the surrounding same-color pixel value and the luminance value, thereby depending on the shooting condition or the like. Since threshold values corresponding to the characteristics of different images are obtained, it is possible to realize a pixel defect correction device that detects and corrects pixel defects without erroneous detection from a dark part to a bright part and from a flat part to an edge part of the image. It has an excellent effect.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a block diagram showing a configuration of a pixel defect correction apparatus 10 according to the first embodiment of the present invention. As shown in FIG. 1, the pixel defect correction apparatus 10 includes an image signal input unit 11, a surrounding same color pixel maximum value calculation unit 12, a surrounding same color pixel minimum value calculation unit 13, a surrounding luminance value calculation unit 14, a threshold value calculation unit 15, A comparator 16, a selector 17, and an image signal output unit 18 are provided. The threshold calculation unit 15 includes an adder, a subtractor, and an integrator that calculate signals output from the surrounding same color pixel maximum value calculation unit 12, the surrounding same color pixel minimum value calculation unit 13, and the surrounding luminance value calculation unit. The

  The image signal input unit 11 sequentially inputs signals in pixel units. Since a signal of a plurality of lines is necessary to obtain a signal of surrounding pixels, a plurality of columns of image signals are input in parallel using a line memory or the like. Here, an example in which five lines are used is described, but the number of lines is not limited to five.

  In the surrounding same color pixel maximum value calculation unit 12 and the surrounding same color pixel minimum value calculation unit 13, the maximum value S <b> 1 and the minimum value S <b> 2 of signals of the same color as the pixel to be inspected are obtained from a plurality of types of color signals arranged in a checkered pattern or a grid pattern. Ask for. At this time, the surrounding same color pixel maximum value calculation unit 12 and the surrounding same color pixel minimum value calculation unit 13 do not include the value of the inspection target pixel in the comparison of the maximum value or the minimum value.

  The ambient brightness value calculation unit 14 calculates a brightness value S4 of pixels around the inspection target pixel. In the complementary color signal processing, the luminance value can be easily obtained from the sum of the signals of adjacent pixels. At this time, the ambient luminance value calculation unit 14 does not include the value of the inspection target pixel in the calculation target.

The difference between the maximum value S1 and the minimum value S2 of the same color signal obtained above is set as a change amount S5 of the surrounding same color pixel value. The change amount offset value S6 is subtracted from the change amount of the surrounding same color pixel value, and then the change amount gain value S7 is integrated. Further, the luminance value offset value S8 is subtracted from the luminance S4 around the inspection target pixel, and then the luminance value gain value S9 is integrated. The base offset value S10 is added to the sum of these to obtain the determination threshold value S11. That is, the determination threshold value S11 is calculated by the following equation (1).
Determination threshold value S11 = {(maximum value S1−minimum value S2 of same color signal) −change amount offset value S6} × change amount gain value S7 + (luminance value S4−luminance value offset value S8 around the inspection target pixel) × luminance value Gain value S9 + Base offset value S10 (1)

  In this way, by subtracting the change amount offset value S6 from the change amount of the surrounding same-color pixel value, and accumulating the change amount gain value S7, it is possible to adjust the change amount of the determination threshold S11 with respect to the image change. Further, by subtracting the luminance value offset value S8 from the surrounding luminance value S4 and then integrating the luminance value gain value S9, the amount of change in the determination threshold S11 with respect to the brightness of the image can be adjusted.

  Next, a value obtained by adding the determination threshold S11 to the maximum value S1 of the same color signal, a value obtained by subtracting the determination threshold S11 from the minimum value S2, and the signal value S3 of the inspection target pixel are input to the comparator 16. The comparator 16 detects a defective pixel using the input value. Specifically, when the value S3 of the pixel to be inspected is larger than the value obtained by adding the maximum value S1 of the same color signal and the determination threshold value S11, it is determined that the pixel to be inspected is a defective pixel. Even when the value S3 of the pixel to be inspected is smaller than the value obtained by subtracting the determination threshold S11 from the minimum value S2 of the same color signal, the pixel to be inspected is determined to be a defective pixel. The comparator 16 inputs the determination result to the selector 17.

The selector 17 uses the result output from the comparator 16 to determine whether or not the inspection target pixel is a defective pixel, and selects an output signal based on the determination result. When the pixel to be inspected is a defective pixel, the value of the pixel to be inspected is replaced with the maximum value S1 or the minimum value S2 of the same color signal. Specifically, the selector 17 passes the pixel value selected based on the following equation (2) to the image signal output unit 18.
When the value of the inspection target pixel> the maximum value of the same color signal + the determination threshold: the maximum value of the same color pixel. The value of the inspection target pixel <the minimum value of the same color signal−when the determination threshold is satisfied: the minimum value of the same color pixel. Target pixel value (2)

  That is, if the value of the inspection target pixel is larger than the sum of the maximum value of the same color signal and the determination threshold value, or if the value of the inspection target pixel is smaller than the difference between the minimum value of the same color signal and the determination threshold value, Is a defective pixel, the selector 17 passes the corrected pixel value to the image signal output unit 18. In cases other than the above, the inspection target pixel is not a defective pixel, and the selector 17 passes the value of the inspection target pixel to the image signal output unit 18.

  The image signal output unit 18 outputs the signal passed from the selector 17 as a pixel value to be inspected.

  Since the pixel defect correction apparatus 10 according to the present embodiment uses a threshold value calculated based on an image to be corrected instead of using a preset threshold value in order to calculate a defective pixel, Accordingly, a defective pixel can be appropriately detected.

  That is, since the threshold value for detecting the pixel defect is obtained by the equation (1), the luminance value of the surrounding pixels becomes small in the dark part of the screen, so the determination threshold value S11 becomes small. Thus, pixel defects can be detected even when the signal level difference between normal pixels and pixel defects is small in an image obtained by copying a dark subject and applying gain. In addition, the determination threshold value is increased in a bright image and the determination threshold value S11 is increased in an edge portion of the image, so that it is difficult to erroneously detect an edge of the image. In addition, in an image with many changes, the amount of change in Expression (1) for obtaining the determination threshold value S11 is large, so that the determination threshold value S11 is large and it is difficult to erroneously detect an edge of the image. As described above, an appropriate threshold value is calculated according to the characteristics of the video, so that defective pixels can be detected and corrected with high accuracy.

  Next, the pixel defect correction apparatus 10 according to the second embodiment of the present invention will be described. The pixel defect correction apparatus 10 of the second embodiment has basically the same configuration as the pixel defect correction apparatus 10 of the first embodiment, but the pixel defect correction apparatus 10 of the second embodiment The difference is that a pixel defect having a large influence is obtained in advance and its position information is recorded. In the following description, a different part from 1st Embodiment is demonstrated.

  FIG. 2 is a block diagram illustrating a configuration of the pixel defect correction apparatus 10 according to the second embodiment. As shown in FIG. 2, in addition to the configuration of the pixel defect correction apparatus 10 of the first embodiment, the pixel defect correction apparatus 10 of the second embodiment includes the surrounding same color pixel average value calculation processing unit 19 and pixels. A defect number counter 20, an address table 21, an address counter 22, a comparator 23, and a selector 24 are provided.

  With these configurations, the coordinates of the defective pixel can be recorded in the address table 21, and the pixel defect can be corrected based on the recorded coordinate information. The operation will be described below.

  First, the operation in the measurement mode in which the coordinates of the defective pixel are recorded in the address table 21 will be described. First, the mode is set to measure the number of defective pixels and their coordinates, and the lens is closed or a subject with a uniform screen is imaged. The values of the pixel defect number counter 20 and the address table 21 are cleared, and the address counter 22 is cleared at the head of the image so as to indicate the coordinates of the pixels and is counted up according to the pixels.

  When a pixel defect is detected by the comparator 22, the pixel defect number counter 20 is incremented, and the value of the address counter 22 is set in the address table 21. By performing this for one field or one frame, the number of pixel defects and position information of the solid-state image sensor can be obtained. Then, the pixel defect correction apparatus 10 records the obtained position information in the address table 21. Here, when the number of obtained defective pixels is larger than a predetermined number, the pixel defect may be detected again by changing the conditions for detecting the defective pixels. Thereby, the amount of information recorded in the address table 21 can be suppressed.

  Next, the normal operation of the pixel defect correction apparatus 10 will be described. In the normal operation, the values of the address table 21 and the address counter 22 are compared by the comparator 23. If the two values match, it is determined as a defective pixel position, and the output of the selector 24 calculates the average value of surrounding pixels of the same color. The pixel defect is corrected by switching to the average value calculated in the process 19.

  The mode of the pixel defect correction apparatus 10 for measuring the number of defective pixels and their position / normal operation is switched by a microcomputer (not shown), and the contents of the address table 21 are written in a nonvolatile memory (not shown). Thus, it is not necessary to detect the defective pixel position every time by returning to the time of restart. In the normal operation, the defective pixels detected from the input video are also corrected as in the first embodiment.

  Since the pixel defect correction apparatus 10 according to the second embodiment records the position of the defective pixel in advance and performs correction based on the recorded position information of the defective pixel, the pixel defect correction apparatus 10 cannot detect the pixel defect detection based on the image. In addition, pixel defects (for example, in the vicinity of edges) can be detected and corrected with high accuracy.

  Further, by using both the correction from the address table 21 and the correction based on the determination of the defective pixel using the threshold value, it is not necessary to record information on all the defective pixels in the address table 21. The size can be reduced. In addition, the circuit scale can be reduced by using the same circuit for the defective pixel detection in the normal operation and the defective pixel detection in the operation of measuring the defective pixel.

  According to the present invention, it is possible to realize a pixel defect correction device that can detect and correct pixel defects from a dark part to a bright part of an image and from a flat part to an edge part without erroneous detection. Therefore, the present invention is useful as a pixel defect correction device that detects and corrects pixel defects.

1 is a block diagram of a pixel defect correction apparatus according to a first embodiment. The block diagram of the pixel defect correction apparatus of 2nd Embodiment

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Pixel defect correction apparatus 11 Image signal input part 12 Surrounding same color pixel maximum value calculation part 13 Surrounding same color pixel minimum value calculation part 14 Surrounding brightness value calculation part 15 Threshold calculation part 16 Comparator 17 Selector 18 Image signal output part 19 Surrounding same color pixel Average value calculator 20 Pixel defect counter 21 Address table 22 Address counter 23 Comparator 24 Selector

Claims (6)

Video input means for inputting video;
In the video input to the video input means, a change amount calculation means for calculating a change amount of the same color pixel value around the inspection target pixel whether or not it is a defective pixel,
A luminance value calculating means for calculating a luminance value of pixels around the inspection target pixel;
Threshold calculating means for calculating a threshold for detecting defective pixels based on the change amount of the same color pixel value and the luminance value;
A defective pixel detecting means for determining whether the inspection target pixel is a defective pixel using the threshold;
Correction means for correcting the pixel value of the defective pixel detected by the defective pixel detection means;
A pixel defect correction apparatus comprising: Recording means for recording coordinates of defective pixels of the image sensor provided in the video input means,
2. The pixel defect correction apparatus according to claim 1, wherein the correction unit corrects a pixel value of a defective pixel at coordinates recorded in the recording unit.   2. The pixel defect correction apparatus according to claim 1, wherein when the change amount of the same-color pixel value is large, the threshold calculation unit calculates a threshold value that is larger than when the change amount is small.   The pixel defect correction apparatus according to claim 1, wherein when the luminance value is large, the threshold value calculating unit calculates a threshold value that is larger than when the luminance value is small. Video input means for inputting video;
In the video input to the video input means, a change amount calculation means for calculating a change amount of the same color pixel value around the inspection target pixel whether or not it is a defective pixel,
A luminance value calculating means for calculating a luminance value of pixels around the inspection target pixel;
Defective pixel detection means for determining whether the inspection target pixel is a defective pixel based on the change amount of the same color pixel value and the luminance value;
Correction means for correcting the pixel value of the defective pixel detected by the defective pixel detection means;
A pixel defect correction apparatus comprising: A video input step for inputting video;
In the video input in the video input step, a change amount calculating step for calculating a change amount of the same color pixel value around the inspection target pixel whether or not it is a defective pixel;
A luminance value calculating step of calculating a luminance value of pixels around the inspection target pixel;
A threshold value calculating step for calculating a threshold value for detecting a defective pixel based on the change amount of the same color pixel value and the luminance value;
A defective pixel detection step of determining whether the inspection target pixel is a defective pixel using the threshold;
A correction step of correcting the pixel value of the defective pixel detected in the defective pixel detection step;
A pixel defect correction method comprising:

Images