JP2006140654A - Circuit and method for correcting defect detection - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To decide a defect with high precision by eliminating a misdecision on a defective pixel of an imaging device without increasing a circuit scale. <P>SOLUTION: A defect detection portion 2 detects a defect of an object pixel of defect detection in an image signal input to a defect detecting and correcting circuit 60 after being photoelectrically converted by the imaging device 22 and if a misdetection detection pattern detection portion 3 detects the defect detection object pixel signal and a pixel signal of its periphery having pixel signal pattern meeting designated conditions, a multiplexer 5 does not output a pixel signal corrected by a correction value generation portion 4 to a trailing-stage image processing portion 25 instead of the defect detection object pixel signal. Thus, it is possible to decide the defect with high precision by eliminating the misdecision on the defective pixel of the imaging device 22 without increasing the circuit scale. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an imaging apparatus, and more particularly, to a defect detection correction circuit and a defect detection correction method for detecting and correcting a white defect when an image is captured by a solid-state imaging device such as a CCD.

  Conventionally, in a fixed imaging device such as a CCD, a video signal portion having a high signal level is output due to a semiconductor crystal defect, and an image display error called a white defect is prominent because the brightness is higher than the surrounding image and white display is conspicuous. There is a technique for detecting and correcting such white defect.

  In the conventional defect detection correction, a defect detection method by shutter closing has been used. Therefore, a defect is detected when there is a white scratch signal in a state where it is completely shielded from light, so that it is possible to detect the defect with considerable accuracy relatively easily. In this defect detection method by closing the shutter, the address of the defect is held in the memory while the shutter is closed, and correction is performed by matching the addresses of every several frames in order to improve detection accuracy. ing. However, in recent solid-state imaging devices that are increasing in the number of pixels, as the number of defects is increased and the amount of information stored in the memory is increased, a large amount of memory capacity is required, which increases the circuit scale and the cost. I have a problem that leads to. Further, in the low-priced model such as a mobile phone camera and an in-vehicle camera, the above method cannot be applied because there is no shutter.

  Therefore, a real-time defect detection and correction circuit that does not use a memory has been developed. In this method, a defect is dynamically detected from peripheral pixel information of the target pixel within one frame, and pixel correction is performed on the detected defect within the same frame. When this method is used for defect detection and correction, it is possible to detect and correct up to infinity without having a memory, and it is possible to detect and correct defects more quickly than the method of detecting defects across frames. .

However, in the real-time defect detection and correction method, when an image with a high spatial frequency, for example, an image such as a resolution chart, is taken, it must be determined whether or not it is a defect based on input signal information of surrounding pixels. There are many cases where an erroneous defect determination is made depending on an input image or noise. For this reason, there are cases in which the image quality is deteriorated, for example, the image has a false color or disappears partially when characters are copied. In order to prevent this, a method of performing pattern recognition of peripheral pixels in more detail has been proposed.
Japanese Patent Application Laid-Open No. 07-59011

  In the real-time defect detection and correction method that eliminates the above-mentioned erroneous defect determination by pattern recognition of surrounding pixels, it is necessary to cover all patterns in order to increase the accuracy of the defect determination. There has been a problem that the circuit scale of the detection correction circuit is increased.

  The present invention has been devised in view of the above circumstances, and an object of the present invention is to perform defect determination with high accuracy without increasing the circuit scale and without erroneous determination of defective pixels of the image sensor. It is an object of the present invention to provide a defect detection correction circuit and defect detection correction method capable of performing

  In order to achieve the above object, the present invention provides a peripheral image signal generating unit that inputs an image signal photoelectrically converted by an image sensor and generates a defect detection target image signal and its peripheral image signal, and the peripheral image signal generating unit. A defect detection means for detecting whether or not the defect detection target pixel signal in the defect detection target image signal is a defective pixel signal based on the image signal information of the defect, and based on the image signal information from the peripheral image signal generation means Erroneous detection pattern detection means for detecting whether the pixel signal of interest and its surrounding pixel signals have a pixel signal pattern satisfying a predetermined condition, and a pixel signal forming the defect detection target image signal as a normal level image A signal correction unit that replaces the signal with a corrected pixel signal, and detects that the defect detection target pixel signal is a defective pixel signal, and the defect detection target pixel signal And a control means for performing control to output the defect detection target pixel signal without replacing it with the correction pixel signal when it is detected that the peripheral pixel signal has a pixel signal pattern satisfying the predetermined condition. It is characterized by.

In addition, the present invention detects whether or not there is a defective pixel signal in the image signal that is photoelectrically converted by the image sensor, and outputs a defective pixel signal by replacing it with a corrected pixel signal. In the detection correction method, when the defect detection target pixel signal and its surrounding pixel signals have a pixel signal pattern that satisfies a predetermined condition, the defect pixel signal is output without being replaced with the correction pixel signal. And

  As described above, in the present invention, when the defect detection target pixel in the image signal photoelectrically converted by the image sensor is input, the defect detection target pixel signal and the surrounding pixel signal are set to a predetermined value. When a false detection condition having a pixel signal pattern that satisfies the condition is satisfied, the defect determination accuracy can be improved by excluding the pixel signal from the defect correction target. Therefore, since it is not necessary to cover all patterns in order to improve the accuracy of defect determination as in the conventional case, it is possible to dramatically reduce false detection of pixel defects without increasing the circuit scale. Accuracy can be improved.

  Also, the number of appearances of the pixel signal pattern, which is a false detection condition, is counted, and if the counted value exceeds a threshold value, it is determined that the imaging target is a special image, and thereafter, the defect correction function is stopped and the erroneous defect is stopped. Deterioration or destruction of the image due to detection can be prevented.

According to the present invention, when the defect detection target pixel in the image signal photoelectrically converted and input by the image sensor is detected, the defect detection target pixel signal and the surrounding pixel signal are detected as false detection pattern conditions. If this is true, by excluding this pixel signal from the defect correction target, it is possible to perform a highly accurate defect determination without increasing the circuit scale and without erroneously determining a defective pixel of the image sensor.
If the number of occurrences of the false detection pattern exceeds the threshold value, it is determined that a special image signal that is not suitable for pixel defect detection has been input, and thereafter the defect correction function is stopped and the image is deteriorated due to false defect detection. Or destruction can be prevented.

  The purpose of highly accurate defect determination without increasing the circuit scale and eliminating defective pixels of the image sensor is to detect defect detection target pixels in the image signal that is photoelectrically converted by the image sensor and input. When the defect detection is performed, if the defect detection target pixel signal and the surrounding pixel signals meet the false detection pattern condition, this is realized by excluding the pixel signal from the defect correction target.

  FIG. 1 is a block diagram showing a configuration of a defect detection and correction circuit according to the first embodiment of the present invention. The defect detection correction circuit 60 includes a line memory 1, a defect detection unit 2, an erroneous detection pattern detection unit 3, a correction value generation unit 4, and a multiplexer 5. The defect detection and correction circuit 60 is incorporated in a solid-state imaging device, and the solid-state imaging device includes a lens 21, an imaging element 22, a preprocessing unit 23, an A / D conversion unit 24, and a subsequent image processing unit 25.

  Next, the operation of the present embodiment will be described. Light incident from the lens 21 is photoelectrically converted by the image sensor 22, and black level adjustment and gain load are performed by the preprocessing unit 23. Thereafter, the analog image signal is converted into a digital image signal by the AD conversion unit 24 and input to the defect detection correction circuit 60 of the image sensor 22. In the defect detection and correction circuit 60, in order to delay the input image signal in the horizontal scanning period (H), the line memory 1 temporarily holds the image signals of the 1H delay signal and the 2H delay signal, and these defect detection target image signals (1H The peripheral pixel level of the defect detection target image signal (1H delay signal) line is monitored by the delay signal) and the peripheral image signal (the current input signal without delay and the 2H delay signal).

  The defect detection unit 2 uses a defect detection target image line input from the A / D conversion unit 24 and a 1H, 2H delay signal line held in the line memory 1 to detect a defect detection target (attention) pixel (1H delay image). And the surrounding pixel level included in the line are compared, and based on the comparison result, it is detected whether the pixel of interest is a defective pixel, and the detection result is output to the multiplexer 5. The false detection pattern detection unit 3 uses the defect detection target image line input from the A / D conversion unit 24 and the 1H and 2H delay signal lines held in the line memory 1 so that the defect detection unit 2 selects the target pixel. It is detected whether or not the erroneous detection pattern is erroneously detected, and the detection result is output to the multiplexer 5. Further, the correction value generation unit 4 receives the defect detection target image line (1H delay signal line), generates a correction pixel signal in which the white point and the black point included in this line are replaced with the signal level of the normal pixel, Output to the multiplexer 5.

  The multiplexer 5 receives the 1H delay signal line (defect detection target signal line) and outputs the pixel signal of the 1H delay signal line to the subsequent image processing unit 25 as it is, or receives the correction pixel signal from the correction value generation unit 4. Whether to output to the subsequent image processing unit 25 is selected based on the detection results of the defect detection unit 2 and the false detection pattern detection unit 3. That is, when the pixel is not defective, the pixel signal of the defect detection target signal line (1H delay signal line) is output to the subsequent image processing unit 25 as it is. If the pixel has a defect but an erroneous detection pattern is detected, the pixel signal of the defect detection target signal line line is output to the subsequent image processing unit 25 as it is. When the pixel is defective and no erroneous detection pattern is detected, the corrected image signal is output to the subsequent image processing unit 25 instead of the defect detection target signal.

  FIG. 2 is a flowchart showing the operation of the defect detection correction circuit 60 of the image sensor shown in FIG. First, in step 101, the defect detection unit 2 performs a process of detecting a pixel defect of the image sensor 22, and when a pixel defect is detected, the process proceeds to step 102. At the same time, processing for detecting a pixel pattern in which detection of a pixel defect is erroneously detected is performed by the erroneous detection pattern detection unit 3, and pixel correction processing is performed by the correction value generation unit 4. In step 102, it is determined whether or not a false detection pattern is detected, and if it is detected, the process ends without performing pixel correction processing. Not performing the correction process is equivalent to outputting the defect detection target pixel signal input from the multiplexer 5 to the subsequent stage as it is. On the other hand, if no erroneous detection pattern is detected, the pixel correction process is performed in step 103 and the process is terminated. This correction processing is equivalent to outputting the corrected pixel signal to the subsequent stage in place of the input defect detection target pixel signal by the multiplexer 5.

  Next, an example of an image including a false detection pattern and a method for removing the same will be described in detail when performing the process of the false detection pattern detection unit 3. FIG. 3 is a diagram schematically illustrating an example of signal levels of peripheral pixels having a color filter of the same color as the pixel of interest (defect detection target pixel). As one of the methods for performing real-time defect detection, there is a method for detecting a defect when the signal difference between a target pixel and a peripheral pixel having a color filter of the same color exceeds or falls below a certain threshold. is there. In that case, when an image signal as shown in FIG. 3 enters the defect detection unit 2, it is not actually a defect but is regarded as a defect because the signal level is significantly lower than that of surrounding pixels.

  Therefore, an example in which an exclusion pattern for preventing such erroneous detection will be shown. Details of the pattern definition will be described with reference to FIG. When an image as shown in FIG. 4B is input to the primary color filter sensor as shown in FIG. 4A, “1” and “8”, “3” and “6” shown in FIG. "2", "4", "5", and "7" signal levels are not relatively different. However, the signal level of “T”, which is the target pixel, is significantly smaller than the other pixels. Also in the example of FIG. 4C, the signal levels of “2” and “7”, “4” and “5” are substantially equal, and “1”, “3”, “6”, “8”. The signal level of “T” is not so different, but the signal level of “T” is significantly different from other pixels.

  When an image of such a certain frequency is input, it is erroneously detected as a defect (in this case, it is regarded as a black spot defect, but the case where the signal amount of “T” is extremely large is considered as well, and this is a white spot defect. The image is destroyed because the correction is performed using peripheral pixels that are greatly different from the actual pixel level.

Therefore, in FIG. 4A, the signal level of the pixel “1” and the pixel “8” is expressed as (1, 8), and the words are defined as follows.
Flat: When there is no difference in signal level between two pixels and it is visually determined that the subject has the same color and brightness.
Separated: When there is a difference in signal level between two pixels, and it is visually determined that an object of different color and brightness is being projected. It should be noted that the degree to which the level differs is determined to be visually different depending on the parameter to be set.
At this time,
Oblique condition A: (1, 8) is flat and (3, 6) is away from (1, 8).
B: (3, 6) is flat and (1, 8) is away from (3, 6).
age,
(1) When A, (3, 6) is larger or smaller than (1, 8).
(2) When B, (1,8) is larger or smaller than (3,6).
In this case, the defect is not detected.

Similarly, in FIG. 4A, the vertical and horizontal conditions C: (4, 5) are flat and (2, 7) are away from (4, 5) in the vertical and horizontal directions of the pixel array.
D: (2,7) is flat and (4,5) is away from (2,7).
age,
(3) When C, (4, 5) is larger or smaller than (2, 7).
(4) At D, (2,7) is larger or smaller than (4,5).
In this case, the defect is not detected.
That is, when any of (1), (2), (3), and (4) is established, it is recognized as a high-frequency pattern and excluded from the defect detection target.

  FIG. 5 is a diagram showing the relationship between the pixel signal level and the threshold value. In order to determine the exclusion condition shown above, a threshold value 1 for determining that it is flat and a threshold value 2 for determining that it is separated are used. The threshold value 1 is calculated and set in consideration of a signal level such as pixel noise obtained experimentally, and is used to give a width to a range in which an object of the same color is shown and that there is no luminance difference. When the noise level of the pixel increases, the width is widened by increasing the threshold value 1. Conversely, when there is little noise, the threshold value 1 is decreased and the range is narrowed. By doing so, judgment errors due to noise are prevented. Threshold 2 is a signal that has a large difference in signal level in the same color filter, a strong contrast, a visually large luminance difference, and a subject of a different color. Level-threshold 1 |> | threshold 2 |> | threshold 1 | is set.

  6A and 6B are diagrams showing the effect. FIG. 6A shows a resolution chart image when the exclusion function is not used, and FIG. 6B shows a resolution chart when the exclusion function is used. As is apparent from the figure, it can be seen that the resolution chart when using the exclusion function is a beautiful image with no destroyed part.

  According to the present embodiment, even when a defect of the target pixel is erroneously detected by the defect detection unit 2, if the erroneous detection pattern is detected by the erroneous detection pattern detection unit 3, the target pixel is not corrected and remains unchanged. Since it is output to the subsequent image processing unit 25, it is possible to drastically reduce false detection without covering all the peripheral pixel patterns for defect determination held in the defect detection unit 2, that is, without increasing the circuit scale. It is possible to perform accurate pixel defects and accompanying pixel correction.

  In addition, by detecting an erroneous detection pattern by the erroneous detection pattern detection unit 3, it is possible to automatically perform dynamic erroneous detection according to the subject, thereby performing defect detection and correction control.

  In addition, since the accuracy of real-time defect detection can be increased without increasing the circuit scale, real-time defect detection can be applied to applications that do not have a mechanical shutter, eliminating the need for factory adjustment. The manufacturing cost can be suppressed.

  FIG. 7 is a block diagram showing a configuration of a defect detection correction circuit of the imaging apparatus according to the second embodiment of the present invention. However, the same parts as those in the first embodiment will be described with the same reference numerals. The defect detection correction circuit 80 includes a line memory 1, a defect detection unit 2, an erroneous detection pattern detection unit 3, a correction value generation unit 4, a multiplexer 5, a counter 6, and a comparator 7. The defect detection and correction circuit 80 is incorporated in a solid-state imaging device, and the solid-state imaging device includes a lens 21, an imaging element 22, a preprocessing unit 23, an A / D conversion unit 24, and a subsequent image processing unit 25.

  Next, the operation of the present embodiment will be described. The operation of the defect detection and correction circuit 80 is almost the same as that of the defect detection and correction circuit 60 shown in FIG. The difference is that the counter 6 counts the number of erroneous detection patterns detected by the erroneous detection pattern detection unit 3 and outputs the count value to the comparator 7. When the input count value exceeds the threshold value, the comparator 7 outputs a signal to notify the multiplexer 5 and stops the operations of the defect detection unit 2, the false detection pattern detection unit 3, and the correction value generation unit 4. Output a signal. Accordingly, the multiplexer 5 thereafter outputs the defect detection target image signal to the subsequent image processing unit 25 as it is. Further, the defect detection unit 2, the false detection pattern detection unit 3, and the correction value generation unit 4 stop their operations to save power.

  FIG. 8 is a flowchart showing the operation of the defect detection and correction circuit shown in FIG. First, in step 201, it is determined whether or not the count value of the counter 6 exceeds the threshold value by the comparator 7. If it exceeds, the defect detection processing is not performed, and the defect detection target image signal is directly sent from the multiplexer 25. The data is input to the subsequent image processing unit 25. If the count value of the counter 6 does not exceed the threshold value, in step 202, the defect detection unit 2 performs a process for detecting a pixel defect in the image sensor 22, and if a pixel defect is detected, the process proceeds to step 203. move on. At the same time, a process for detecting a pattern in which detection of a pixel defect is erroneously detected is performed by the erroneous detection pattern detection unit 3, and a pixel correction process is performed by the correction value generation unit 4.

In step 203, it is determined whether or not an erroneous detection pattern is detected. If it is detected, the count value of the counter 6 is incremented by +1 in step 205, and then the process is terminated without performing the pixel correction process. Not performing the correction process is equivalent to outputting the defect detection target pixel signal to the subsequent stage as it is by the multiplexer 5. On the other hand, if no erroneous detection pattern is detected, the pixel correction process is performed in step 204 and the process is terminated. This correction processing is equivalent to outputting the corrected pixel signal to the subsequent stage in place of the defect detection target pixel signal by the multiplexer 5. Since such processing is performed during the operation of the imaging apparatus, when a special image containing a large amount of high-frequency components is taken, the number of erroneous detection patterns detected is large, so the count value of the counter 6 is immediately set to the threshold value. Therefore, the mode is switched to a mode in which pixel defect detection and correction are not performed, and the captured image can be prevented from being destroyed by incorrect correction.

  According to the present embodiment, when the number of erroneous detection patterns detected at the time of image capturing is large, it is determined that a special image containing a large amount of high-frequency components has been captured, and the pixel defect detection and correction mode is not performed. By doing so, it is possible to prevent the captured image from being deteriorated or destroyed by incorrect correction. Other effects are the same as those of the first embodiment.

  In addition, this invention is not limited to the said embodiment, In the range which does not deviate from the summary, it can implement also with another various form in a concrete structure, a function, an effect | action, and an effect. The same effect can be obtained by applying the present invention to any apparatus as long as it is an apparatus that processes an image captured by an imaging device.

1 is a block diagram showing a configuration of a solid-state imaging device equipped with a defect detection correction circuit according to a first embodiment of the present invention. 3 is a flowchart for explaining an operation process of the defect detection and correction circuit shown in FIG. 1. It is the figure which showed typically an example of the signal level of a surrounding pixel which has a color filter of the same color as the pixel of interest (defect detection object pixel). It is a figure explaining the reason for the defect defect detection of the pixel by the defect detection part shown in FIG. It is a figure for demonstrating the conditions of the pixel pattern which the defect detection part shown in FIG. FIG. 2 is a diagram comparing an image when control is performed to determine whether or not to perform pixel correction by a false detection pattern when a pixel defect is detected by the defect detection and correction circuit illustrated in FIG. 1 and when the pixel correction is not performed. It is the block diagram which showed the structure of the solid-state imaging device carrying the defect detection correction circuit based on the 2nd Embodiment of this invention. It is a flowchart explaining the operation | movement process of the defect detection correction circuit shown in FIG.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Line memory, 2 ... Defect detection part, 3 ... False detection pattern detection part, 4 ... Correction value generation part, 5 ... Multiplexer, 6 ... Counter, 7 ... Comparator, 21 ... Lens , 22... Image sensor, 23... Pre-processing unit, 24... A / D conversion unit, 25.

Claims (6)

Peripheral image signal generation means for inputting an image signal photoelectrically converted by the image sensor and generating a defect detection target image signal and its peripheral image signal;
Defect detection means for detecting whether or not a defect detection target pixel signal in the defect detection target image signal is a defective pixel signal based on image signal information from the peripheral image signal generation means;
A false detection pattern detection means for detecting whether or not the defect detection target pixel signal and the peripheral pixel signal have a pixel signal pattern satisfying a predetermined condition based on image signal information from the peripheral image signal generation means;
A signal correcting means for replacing the pixel signal forming the image signal with a normal level image signal to obtain a corrected pixel signal;
When the defect detection target pixel signal is detected to be a defective pixel signal, and the defect detection target pixel signal and its peripheral pixel signal are detected to have a pixel signal pattern that satisfies the predetermined condition, the defect detection is performed. A defect detection / correction circuit comprising: control means for performing control to output a target pixel signal without replacing it with the correction pixel signal. Counting means for counting the number of times of detection of the pixel signal pattern that satisfies the predetermined condition by the erroneous detection pattern detection means;
Comparing means for comparing the count value of the counting means with a preset threshold value,
2. The defect detection and correction circuit according to claim 1, wherein the control means outputs the input image signal thereafter when the count value exceeds the threshold value.   The predetermined condition is that when the range of the defect detection target pixel signal and its peripheral pixel signal is rectangular, the pixel signals on two diagonal lines have relatively close levels, and the pixel signals on one diagonal line When the level is farther than the level of the pixel signal on the other diagonal line, or the pixel signals on the two vertical and horizontal lines are relatively close to each other, and the level of the pixel signal on one line is the other 2. The defect detection / correction circuit according to claim 1, wherein the defect detection / correction circuit is separated from a level of a pixel signal on the line.   3. The defect detection / correction circuit according to claim 2, wherein when the count value exceeds the threshold value, operation of a circuit related to defect detection / correction of the pixel signal is stopped. A defect detection and correction method that detects whether or not there is a defective pixel signal in an image signal that is photoelectrically converted by an image sensor and replaces it with a corrected pixel signal if there is a defective pixel signal. ,
When the defect detection target pixel signal and its surrounding pixel signals have a pixel signal pattern that satisfies a predetermined condition, even if the defect detection target pixel signal is detected as a defective pixel signal, this is detected as the correction pixel signal. A defect detection and correction method characterized in that the defect is output without being replaced.   6. The defect detection according to claim 5, wherein the number of appearances of the pixel signal pattern satisfying the predetermined condition is counted, and if the counted value exceeds a predetermined threshold value, the input image signal is output thereafter. Correction method.

Images