JPH0730905A - Pixel defect correction device - Google Patents

Abstract

(57) [Abstract] [Purpose] Even if there is a flaw in a highly saturated background, this flaw is detected and corrected to prevent deterioration of the image quality due to the flaw. [Configuration] LPFs 4, 5 and 6 for extracting an average saturation level from R, G and B signals read from a plurality of solid-state image pickup devices 1, 2 and 3, and R, G and B signals LPF
Has subtracters 7, 8, 9 for subtracting the outputs of 4, 5, 6;
Detection circuit 21 which receives the outputs of the subtracters 7, 8 and 9
The position of the flaw is detected by the correction circuit 22 and the output of the detection circuit 21 corrects the flaw portion of the R, G, B signals. Even if there is a flaw in the highly saturated background, the influence of the background saturation is exerted. Control to eliminate.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pixel defect correcting apparatus for detecting and correcting a pixel defect existing in a solid-state image sensor in an image sensor using a solid-state image sensor such as CCD.

[0002]

2. Description of the Related Art Generally, it is known that in a solid-state image pickup device formed of a semiconductor such as CCD, image quality is deteriorated due to local crystal defects of the semiconductor. An image defect in which a constant bias voltage is always added to the imaging output according to the amount of incident light appears as a high-intensity white dot on the monitor screen if this image defect signal is processed as it is. There is. In addition, those having low photosensitivity appear as black dots and are called black scratches (hereinafter, pixel defects are called scratches).

Conventionally, the detection of scratches having a plurality of solid-state image pickup devices is disclosed in, for example, Japanese Patent Application Laid-Open No. 56-149181. In this method, the difference output of each solid-state image sensor is taken out, and if the value is a certain value or more, it is detected as a flaw.

A concrete structure is shown in FIG. In FIG.
Reference numerals 2 and 3 denote solid-state image pickup devices that output R (red), G (green), and B (blue), and 21 are detection circuits that detect flaws in the R, G, and B signals read from the solid-state image pickup device. Reference numeral 22 is a correction circuit for correcting the scratched portion of the R, G, B signals based on the detection output of the detection circuit 21, and 23, 24, 25 are output terminals for outputting the corrected R, G, B signals. .

FIG. 9 shows an example of the configuration of the detection circuit 21 shown in FIG. 101 and 102 are subtractors for subtracting R and B inputs from G inputs, 103 and 104 are comparators for outputting "1" if the outputs of the subtracters 101 and 102 are equal to or more than a certain value, 105
Reference numeral 107 denotes a protrusion amount detection circuit that outputs "1" when the pixel of interest is larger or smaller than a peripheral pixel by a certain amount or more. Reference numeral 108 indicates that either or both of the outputs of the comparators 103 and 104 are "1". And the protrusion amount detection circuit 105
When the detection output of ~ 107 is "1", it is an arithmetic unit that determines the pixel as a flaw.

Usually, since the output signals obtained from the solid-state image pickup device are almost the same, the outputs of the subtractors 101 and 102 are small, and when these are compared with a constant value by the comparators 103 and 104, both outputs are constant. It becomes "0".

On the other hand, when only the G solid-state image sensor has a flaw, the input to the detection circuit 21 of FIG. 8 is as shown in FIG.
As in (b), only the scratched portion of G has a high level. Therefore, the outputs of the subtractors 101 and 102 in FIG. 9 have high levels only in the scratched portions as shown in (c) and (d) of FIG.
When compared with a constant value in 04, the output is "1" only in the scratched portion. On the other hand, the G protrusion amount detection circuit 10 of FIG.
In 6 as well, the output is "as shown in FIG.
Since it is 1 ″, the calculator 108 of FIG. 9 outputs a signal representing the flaw position of G to the correction circuit 22 of FIG. 8 as shown in (f) of FIG.

The protrusion amount detection circuits 105 to 10 shown in FIG.
11 shows an example of the configuration of FIG. 7 and FIG. 12 shows an example of the arithmetic unit 108 of FIG.

Regarding the correction of pixel defects, Japanese Patent Laid-Open No. 62-
Several methods are shown in the 8666 publication. For example, a method of substituting one pixel or two pixels before, a method of substituting the average of previous and next pixel values, or a method of thinking in the vertical direction and substituting by one pixel above, an average of upper and lower pixel values There is a method of replacing with.

Here, it is assumed that the correction circuit substitutes the average of the pixel values before and after, and the block diagram is as shown in FIG. 13, and the operation is as follows. The input signal passes through the delay circuits 141 and 142 and extracts the value of the pixel of interest in the center and the pixel values before and after it. An average value of these pixel values before and after the pixel of interest is calculated and used as a correction signal. According to the detection output of the detection circuit, normally, the value of the pixel of interest in the center is output, and when it is determined that there is a flaw, a correction signal is output.

As described above, by comparing the difference output between a plurality of solid-state image pickup devices, a pixel protruding by a certain level or more with respect to the values of the surrounding pixels can be detected as a flaw and is inconspicuous. Can be corrected.

[0012]

However, with the defect correction device having the above-described structure, if there is a defect in the high-saturation background color, it is impossible to detect this defect.

An example will be described with reference to FIG. For example, when the background color is Mg (magenta: R and B are at a high level) and there are G scratches therein, the outputs of the solid-state imaging devices 1, 2, and 3 in FIG. 8 are (a) and (b) in FIG. ), G has a high level only in the scratched portion, and R and B are always high level signals. Therefore, the output of the protrusion amount detection circuit 106 of G in FIG. 9 is “1” as in (e) of FIG. 14 as in the case of low saturation, but the outputs of the subtractors 101 and 102 in FIG. As in (c) and (d) of 14, the scratched portion does not have a high level, and when this is compared with a constant value by the comparators 103 and 104 of FIG. 9, the output of the scratched portion is “0”. Becomes For this reason, it becomes impossible to detect a flaw by the output of the arithmetic unit 108 in FIG.

As described above, when there is a flaw in the high-saturation background, the flaw cannot be detected and cannot be corrected, so that there is a problem that the flaw deteriorates the image quality. There is.

Further, in the defect correction circuit having the above structure, in the case of a high saturation color high frequency signal, this high frequency signal is erroneously detected as a defect.

An example will be described with reference to FIG. For example, in the case of a high-frequency signal of a high saturation color such as a green grove, the output of the solid-state imaging device 2 of G in FIG. 8 is a high-frequency signal as shown in FIG. On the other hand, the outputs of the solid-state image pickup devices 1 and 3 of R and B in FIG. 8 are low level signals as shown in FIG. Therefore, the outputs of the subtractors 101 and 102 in FIG. 9 have a high level in the high level part of G as shown in (c) and (d) of FIG. 15, and these are compared by the comparators 103 and 104 in FIG. In that case, the output of the high level section becomes "1". On the other hand, G in FIG.
Also in the protrusion amount detection circuit 106, the high level portion of G becomes "1", and this portion is erroneously detected as a flaw.

As described above, when there is a high saturation color high frequency signal, this high frequency signal is erroneously detected as a flaw,
Since the correction circuit erroneously corrects as shown in FIG. 15 (g), there is a problem that the high frequency portion is eliminated and the image quality is deteriorated.

[0018]

In order to solve the above-mentioned problems, the present invention, in addition to the conventional method, adds an average saturation from R, G, B signals read from a plurality of solid-state image pickup devices. It has an LPF for extracting the level and a subtractor for subtracting the LPF output from the R, G, B signals read from the solid-state image sensor, and the flaw detection is performed by a detection circuit having the output of the subtractor as an input. The correction circuit corrects the scratched portion of the R, G, B signals by the detection circuit output.

In order to solve the above-mentioned problems, the present invention extracts an average saturation level from R, G, B signals read from a plurality of solid-state image pickup devices in addition to the conventional method. And a detection circuit for extracting a saturation level from the LPF output. The conventional flaw detection circuit output is controlled by the saturation level detection output to detect the position of the flaw, and R, It has a configuration including a correction circuit for correcting the scratched portion of the G and B signals.

[0020]

According to the present invention, with the above configuration, the average saturation level is extracted from the R, G, B signals by the LPF, and the LPF is extracted.
The influence of the background color is eliminated by subtracting the F output from the R, G, B signals. Therefore, even if there is a flaw in the highly saturated background, it is possible to detect and correct this flaw, and it is possible to prevent deterioration of image quality due to the flaw.

Further, the present invention has the above-mentioned structure.
An average saturation level is extracted by the LPF from the R, G, B signals, and the saturation level of the background is detected based on the LPF output. For example, when this detection level is low, it is determined to be a low-saturation portion, and the flaw detection circuit performs conventional detection. On the other hand, when this detection level is high, it is determined to be a high-saturation portion, detection of the flaw detection circuit is stopped, or the detection reference is changed so that the high-frequency signal in the high-saturation portion is not detected as a flaw. As described above, even when there is a high-saturation color high-frequency signal, detection of scratches in this part is stopped, or
By changing the standard, false detection and correction are eliminated,
Prevents deterioration of image quality.

[0022]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of a first embodiment of the present invention.
Shown in. In FIG. 1, reference numerals 1, 2, and 3 denote solid-state image pickup devices that output R (red), G (green), and B (blue) signals.
5, 5 and 6 are R read from the solid-state image pickup devices 1, 2 and 3,
LPF for detecting an average saturation level from G and B signals,
Reference numerals 7, 8 and 9 are subtractors for subtracting the outputs of LPFs 4, 5 and 6 from R, G and B signals read from the solid-state image pickup devices 1, 2 and 3, and 21 is a subtractor for each of R, G and B. A detection circuit for detecting a flaw from the output signal, a correction circuit 22 for correcting a flaw portion of the R, G, B signals based on the detection output of the detection circuit 21, and 23, 24, 25 for the corrected R, G , B signals are output terminals.

The operation of the pixel defect correction device constructed as above will be described with reference to FIG.

For example, if the background color is Mg (magenta: R, B)
Is high level) and there are G scratches therein, the outputs of the solid-state imaging devices 1, 2 and 3 in FIG. 1 are as shown in (a) and (b) of FIG. At high level, R and B are always high level signals. On the other hand, LPF4 of FIG.
By extracting only the low frequency components of the R, G and B signals at 5 and 6, G becomes a low level signal and R and B become a high level signal as shown in (c) and (d) of FIG. . This LPF4
The outputs of 5 and 6 are converted into R, R by the subtractors 7, 8 and 9 of FIG.
2E, in which the influence of saturation is removed by subtracting from the outputs of the G and B solid-state image pickup devices 1, 2, and 3, respectively.
Obtain the output of (f).

By thus inputting the signal from which the influence of the saturation is removed to the detection circuit 21 of FIG. 1, even when the background saturation is high, it is treated in the same manner as a flaw when the background saturation is low. It becomes possible.

When the background has low saturation and there is a flaw of G, the outputs of the LPFs 4, 5 and 6 of FIG. 1 are low level signals, so that the subtractors 7 and 8 of FIG. , 9 is not affected, and the flaw can be detected as in the conventional case.

The detection circuit 21 of FIG. 1 is constructed as shown in FIG. 9 as in the conventional case, and outputs a G flaw correction signal to the correction circuit 22 of FIG. 1 in order to correct the G flaw by the above input.

The correction circuit 22 shown in FIG. 1 can also be configured as shown in FIG. 13 as in the conventional case.

The second embodiment of the present invention will be described below with reference to the drawings. A block diagram of the second embodiment of the present invention is shown in FIG. In FIG. 3, 1, 2, and 3 are R (red) and G
Solid-state image sensor that outputs (green) and B (blue) signals,
4, 5, 6 are R, G, B read from the solid-state image sensor
LPF for detecting an average saturation level from the signal, 10 is a saturation level creating circuit for extracting the saturation level from the outputs of LPFs 4, 5 and 6, and 21 is read from the solid-state image pickup elements 1, 2 and 3. A flaw detection circuit for detecting flaws from the R, G, B signals according to the saturation level, and a correction circuit 22 for correcting flaw portions of the R, G, B signals based on the detection output of the flaw detection circuit 21. , 23, 24, 25 are corrected R, G, B
It is an output terminal for outputting a signal. The operation of the pixel defect correction device configured as described above will be described with reference to FIG.

For example, in the case of a high-saturation color signal of a high saturation color such as a green grove, the outputs of the solid-state image pickup devices 1, 2 and 3 in FIG. 3 are such that G is a high-frequency signal as shown in FIGS. Next to
R and B are low level signals.

On the other hand, LPFs 4, 5 and 6 of FIG.
By extracting only the low-frequency component of the B signal, the output is G as a high-level signal, as shown in (c) and (d) of FIG.
R and B are low level signals. This LPF 4, 5, 6
Is processed by the saturation level generating circuit 10 in FIG.

FIG. 5 is a diagram showing an example of the internal structure of the saturation level generating circuit 10 of FIG. 3, and 31 is the LP of the R, G, B signals.
A circuit for detecting the maximum value of the F output, a circuit for detecting the minimum value 32, and a subtracter 33 for subtracting the minimum value from the maximum value. In this way, a signal having a level corresponding to the saturation of the background is output.

Therefore, in the case of a high-saturation high-frequency signal such as a green grove, the output of the saturation level generation circuit 10 in FIG. 3 becomes a high level as shown in FIG. 4 (e).

Here, the flaw detection circuit 21 of FIG.
As shown in FIG. 7, the protrusion amount comparison is performed using the fixed value in the related art, and the protrusion amount is compared using, for example, the conventional fixed value plus the output of the saturation level generating circuit. As a result, in the low saturation portion, detection is performed at the same level as before, and in the high saturation portion, the comparison value is increased and the detection level of the protrusion amount is increased.

Therefore, since the protrusion amount detection level in the case of the input signal is large, the high frequency signal is not erroneously detected as a flaw, and the output of the flaw detection circuit 21 of FIG.
As shown in (f) of FIG. 4, the flaw correction signal is not output to the correction circuit 22 of FIG. Therefore, even if there is a high-saturation color high-frequency signal, the output of the correction circuit 22 in FIG.
(G), the high-frequency signal does not malfunction as a flaw and the image quality does not deteriorate.

When the background saturation is low and there is no malfunction of the high frequency signal, the output of the saturation level generation circuit 10 in FIG. 3 is at a low level, and the flaw can be detected at the conventional level.

In the above embodiments, the method of controlling the comparison value of the protrusion amount of the flaw detection circuit according to the saturation level has been described. However, the output of the saturation level creating circuit is compared with a fixed value to obtain "0" or A method in which only "1" is output and the output of the flaw detection circuit is turned off only when the saturation is high is also possible.

Further, the saturation level detection of FIG.
The same can be applied to the case where the G and B signals are converted into color difference signals and then detected by another method such as calculation.

Further, in the above embodiments, the R, G and B signals are applied with the LPF to detect the saturation level. However, in order to reduce the circuit scale, after the color difference signal or the saturation level is set, the LPF is used. It is also possible to apply a configuration that applies a multiplication.

[0041]

As is apparent from the above description, according to the present invention, even if there is a flaw in a high-saturation background, this flaw can be detected and corrected, and deterioration of the image quality due to the flaw can be prevented. It is possible to prevent.

Further, according to the present invention, even if there is a high-saturation color high-frequency signal, the signal and the defect are distinguished from each other, the erroneous detection / correction is eliminated, and only the defect is corrected to prevent the deterioration of the image quality. Is possible.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a pixel defect correction device according to a first embodiment of the present invention.

FIG. 2 is a signal waveform diagram for explaining the operation of the pixel defect correction device in the first embodiment.

FIG. 3 is a block diagram showing a configuration of a pixel defect correction device according to a second embodiment of the present invention.

FIG. 4 is a signal waveform diagram for explaining the operation of the pixel defect correcting apparatus in the second embodiment.

FIG. 5 is a block diagram showing an internal configuration of a saturation level creating circuit according to the second embodiment.

FIG. 6 is a block diagram showing an internal configuration of a flaw detection circuit according to the second embodiment.

FIG. 7 is a block diagram showing an internal configuration of a protrusion amount detection circuit according to the second embodiment.

FIG. 8 is a block diagram showing a configuration of a conventional pixel defect correction device.

FIG. 9 is a block diagram showing an internal configuration of a defect detection circuit in a conventional example.

FIG. 10 is a signal waveform diagram for explaining the operation of a conventional pixel defect correction device.

FIG. 11 is a block diagram showing an internal configuration of a protrusion amount detection circuit in a conventional example.

FIG. 12 is a block diagram showing an internal configuration of an arithmetic unit in a conventional example.

FIG. 13 is a block diagram showing an internal configuration of a conventional flaw correction circuit.

FIG. 14 is a signal waveform diagram for explaining the operation of a conventional pixel defect correction device.

FIG. 15 is a signal waveform diagram for explaining the operation of a conventional pixel defect correction device.

[Explanation of symbols]

 1-3 Solid-state image sensor (CCD) 4-6 LPF 7-9 Subtractor 21 Detection circuit 22 Correction circuit 23-25 Output terminal

Claims (2)

[Claims] 1. A plurality of solid-state image pickup devices, and an LP for inputting a signal read from the solid-state image pickup device.
F and the LPF from the signal read from the solid-state image sensor.
A subtractor for subtracting an output; a detection circuit for detecting a pixel defect of the solid-state image sensor output from the subtractor output; and a correction circuit for correcting a pixel defect part of the solid-state image sensor output by the detection circuit output. A pixel defect correction device characterized by the above. 2. A plurality of solid-state image pickup devices, and an LP for inputting a signal read from the solid-state image pickup device.
F, a saturation level creating circuit that creates a saturation level from the LPF output, a detection circuit that detects a pixel defect of the solid-state image sensor output according to the saturation level creating circuit output, and the detection circuit output Accordingly, the pixel defect correction device is provided with a correction circuit for correcting the pixel defect portion of the output of the solid-state image sensor.