JP2002271805A - Noise elimination circuit - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve a problem that it is impossible to sufficiently eliminate a noise due to a defect in a solid-state image pickup element or the like by only adding a low frequency component of one color to a high frequency component of other color when a change it the signal of the one color is inverse to a change in the signal of the other color in the case of eliminating the noise. SOLUTION: When a change in a signal whose noise is to be eliminated is inverse to a change in a reference signal, a high frequency component processing circuit 3 zeroes a high frequency component of the reference signal or inverts it depending on the change. An adder circuit 4 adds an output of a low-pass filter 1 to an output of the high frequency component processing circuit 3, and a selection circuit 5 selects an input signal GI or an output of the adder circuit 4 on the basis of a noise range designation signal NF.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a noise removing circuit for removing noise of a video signal due to a defect of a photodiode of a solid-state imaging device.

[0002]

2. Description of the Related Art In a video camera using a solid-state image sensor, if a photodiode of the solid-state image sensor is defective,
Depending on the type of the defect, the defective portion appears as a high level point in the image. For example, in an RGB three-panel video camera, if a G image sensor has a photodiode defect, the level of G increases in a portion that should be black,
It may be observed as a green dot on the screen. Here, R, G, and B are red, green, and blue video signals of the three primary colors of light. In the following description, a red video signal is simply an R or R signal, and a green video signal is G or G. The signal and the blue video signal are described as B or B signal.

[0003] With respect to such noise caused by the defect of the photodiode of the solid-state imaging device, an RGB three-panel video camera, as disclosed in Japanese Patent Application Laid-Open No. 9-284793, discloses an image from which noise is removed. By adding the high-frequency component of the video signal of another color to the low-frequency component of the signal,
There is a method for removing noise. This method usually involves R,
G and B are based on the fact that they have a correlation with each other,
As long as there is a correlation between R, G, and B, noise can be made inconspicuous.

However, some artificially generated objects (images) have no correlation among R, G, and B. FIG.
Is an example of a waveform diagram of R, G, and B signals obtained from the solid-state imaging device. The horizontal axis represents sample numbers, and the vertical axis represents amplitude. Signal waveform is obtained by extracting a portion used to describe only six samples, indicating the value of the 10-bit system as an amplitude, i.e. the integer value up to 0-1023. In the following description, the waveform diagrams of the video signals are all unified into this format.

[0005] When an object such as that described above is photographed, FIG.
GS shown in (1), BI shown in FIG. 9 (2), FIG. 9 (3)
It is assumed that a video signal like RI shown in FIG. When such an object is photographed, if there is a defective portion of the solid-state imaging device that outputs the G signal at a change portion of the video signal, the signal becomes a noisy signal as shown by a signal GI in FIG. Would. For such a signal GI,
When trying to remove noise by the method disclosed in Japanese Patent Application Laid-Open No. 9-284793, the signal GI is converted into a signal GO as shown in FIG. 9 (9). The process of such signal processing will be specifically described below.

FIG. 8 is a block diagram showing a configuration example of a conventional noise elimination circuit. Here, a case will be described in which noise of the G signal is removed using the R signal. FIG.
In contrast to the original signal GS shown in (1), in the signal GI actually input, it is assumed that the third and fourth samples are noise. This signal GI is input to the low-pass filter 1 in FIG. 8, and is a signal obtained by averaging the signals two samples before and two samples after the processing sample, that is, FIG.
As shown in (5), the signal GL which is a low-frequency component of the G signal
Is generated.

On the other hand, the signal RI is input to the high-pass filter 2. The high-pass filter 2 converts the processed sample into 2
A signal is generated by subtracting the average value of the signals before and after two samples. This signal is a signal RH representing the high frequency component of the R signal, as shown in FIG. And
The addition circuit 4 adds the signal GL and the signal RH, and
A noise removal signal GC as shown in (7) is generated.

Further, since the noise is a 3-th sample and 4 th sample, 3-th sample and a signal indicating that the 4-th sample is noise, that is, 3-th sample and 4 th sample, as shown in FIG. 9 (8) The noise range designation signal NF which is 1 and 0 otherwise is applied to the selection circuit 5. Selection circuit 5
Selects the noise removal signal GC when the signal NF is 1, selects the input signal GI when the signal NF is 0, and generates an output signal GO as shown in FIG. 9 (9).

Thus, in the conventional noise elimination circuit shown in FIG. 8, a signal containing another noise different from the original noise, such as the output signal GO, is output from the original signal GS.

[0010]

In the above-described conventional noise elimination circuit, the original signal GS shown in FIG.
Since the signal GO is converted into the signal GO shown in (9), especially when the video camera is moved horizontally, when the defect position of the solid-state imaging device and the boundary portion of the subject overlap, something flashes for a moment due to noise. Such a thing may be seen, and the existence of a defect may be recognized.

Even when the angle of view of the video camera is fixed, if the defect position of the image sensor and the boundary portion of the subject overlap, the noise may be seen as flickering at an indefinite cycle.

[0012] The present invention has been made such has been made in view of the conventional problems, R, G, even if there is no correlation to the video signal of B, know that there are defects in the solid-state imaging device It is an object of the present invention to provide a noise elimination circuit that makes it difficult.

[0013]

According to a first aspect of the present invention, there is provided an image pickup apparatus comprising: a pickup device which includes a first image signal and a second image signal which are simultaneously outputted from an image sensor; A noise removal circuit for suppressing the noise component, wherein the low-pass filter extracts a low-frequency component from the first video signal including the noise component; and the second video signal does not include the noise component. from extracting a high-frequency component and a high-pass filter, the first when the polarity of the change amount of the polar and the second video signal of the variation of the video signal is the same is output as the high-frequency component, wherein a high frequency component processing means when the polarity of the change amount of the first video signal of the variation of the polarity and the second video signal is different from that output by attenuating high-frequency components of the high-pass filter, the low Range component and the high range component processing means Adding means for adding the output signals of the first and second video signals, and selecting the output signal from the adding means when the first video signal is noise according to the signal indicating the noise portion; And selecting means for selecting.

According to a second aspect of the present invention, when the image signal including the first and second video signals output simultaneously from the image sensor includes noise inherent in the image sensor, the noise component is suppressed. A low-pass filter for extracting a low-frequency component from the first video signal including the noise component, and a high-frequency component from a second video signal not including the noise component. When the polarity of the amount of change of the first video signal is the same as the polarity of the amount of change of the second video signal, the high-pass filter outputs the high-frequency component as it is, and the change of the first video signal. High-frequency component processing means for inverting and outputting the high-frequency component when the polarity of the amount is different from the polarity of the variation amount of the second video signal; and output from the low-frequency component and the high-frequency component processing means. Adding means for adding a signal; Wherein in accordance with a signal indicating a partial first
And selecting means for selecting the output signal from the adding means when the video signal is noise, and selecting the first video signal when the video signal is not noise.

[0015] According to a third aspect of the present invention, when the image signal including the first, second, and third video signals simultaneously output from the image sensor includes noise inherent in the image sensor,
A noise removal circuit for suppressing the noise component, comprising: a low-pass filter that extracts a low-frequency component from the first video signal containing the noise component; and a second video signal that does not contain the noise component. A first high-pass filter that extracts a first high-frequency component, a second high-pass filter that extracts a second high-frequency component of the third video signal that does not include the noise component, When the polarity of the amount of change of the first video signal is the same as the polarity of the amount of change of the second video signal, the first high-frequency component is selected, and the polarity of the amount of change of the first video signal and the polarity of the Unlike the polarity of the change amount of the second video signal,
When the polarity of the change amount of the first video signal is the same as the polarity of the change amount of the third video signal, the second high-frequency component is selected, and the polarity of the change amount of the first video signal and When the polarity of the amount of change of the second video signal is different, and the polarity of the amount of change of the first video signal is different from the polarity of the amount of change of the third video signal, the first high-frequency component and the High-frequency component processing means for selecting and outputting a signal in which one of the second high-frequency components has been attenuated, and addition means for adding the low-frequency component and an output signal from the high-frequency component processing means Selecting means for selecting an output signal from the adding means when the first video signal is noise according to a signal indicating a noise portion, and selecting the first video signal when the first video signal is not noise. It is characterized by the following.

[0016]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a noise removing circuit according to an embodiment of the present invention will be described with reference to the drawings.

(Embodiment 1) FIG. 1 is a block diagram showing a configuration of a noise removing circuit according to Embodiment 1 of the present invention. The noise elimination circuit is a circuit that suppresses the noise component when an image signal including first, second, and third video signal components output from the solid-state imaging device includes noise inherent in the solid-state imaging device. It is. This noise elimination circuit includes a low-pass filter 1, a high-pass filter 2, a high-frequency component processing circuit 3, an addition circuit 4, and a selection circuit 5. In the following description, it is assumed that the first video signal includes noise due to a defect of a specific photodiode of the solid-state imaging device.

The low-pass filter 1 is a filter for extracting a low-frequency component from a first video signal containing a noise component.
The high-pass filter 2 has a characteristic opposite to that of the low-pass filter 1 and is a filter for extracting a high-frequency component from a second video signal that does not include a noise component. The high-frequency component processing circuit 3 is a circuit that adjusts the amplitude of the high-frequency component output from the high-pass filter 2. The addition circuit 4 is a circuit that adds the output signal of the low-pass filter 1 and the output signal of the high-frequency component processing circuit 3. The selection circuit 5 receives the input video signal and the addition circuit 4
Is a circuit that selects and outputs one of the output signals according to the control signal.

[0019] The noise removing circuit having such a configuration, and the original G signal and the R signal is changing in opposite directions at the output level, when the R signal is also used to remove the noise of the G signal The operation will be described.
FIG. 2 is an explanatory diagram showing signal waveforms at various parts under such conditions. As shown in FIG. 2D, a signal GI in which the third and fourth samples are noise is input as the first video signal to the original signal GS shown in FIG. 2A. I do. The Figure 2 as a second video signal (2)
It is assumed that a signal RI as shown in FIG.

First, when a signal GI containing noise is input to the low-pass filter 1 of FIG. 1, a signal obtained by averaging the signals two samples before and two samples later, that is, as a low-frequency component of the G signal in FIG. (5) signal GL as shown in is generated. On the other hand, when the input signal RI is input to the high-pass filter 2, as a conversion process of the reverse characteristic of the low-pass filter 1, the signal two samples before and two samples after the current signal is added and averaged. A signal RH as shown in FIG. 2C is generated as a signal obtained by subtracting the signal, that is, as a high frequency component of the R signal. The high-frequency component processing circuit 3, compared with 2 samples before the signal, to determine the increase or decrease of 2 samples after the signal, and an increasing number of G signal reduces the R signal, or and increasing R signal decreases is G signal In this case, 0 is output, and otherwise, the signal RH is output as it is. Thus, the high-frequency component processing circuit 3 generates a signal GH as shown in FIG. 2 (6) as a corrected high-frequency component of the G signal. Then, the addition circuit 4 adds the signal GH and the signal GL, and as shown in FIG.
Generating a correction signal GC of the signal.

With respect to the signal GC obtained in this way, FIG.
When the noise range designation signal NF shown in (8) is 1, the selection circuit 5 selects the signal GC, and when the signal NF is 0, the selection circuit 5 selects the input signal GI. As a result, as shown in FIG. 2 (9), a signal GO in which a noise portion is replaced is obtained as an output signal of the G signal. Here, the signal NF is shown in FIG.
As shown in (8), when the third sample and the fourth sample are noise, the value is 1, and the noises are not 1, 2, 5,.
The signal becomes 0 at the time of the sixth sample.

By the above operation, the original signals are R, G,
Noise can be reduced even when there is no correlation between B and the G signal and the R signal are changing in opposite directions.

[0023] The noise removing circuit in the second embodiment of the present invention will now Embodiment 2 will be described with reference to the drawings. The basic configuration of the noise elimination circuit is the same as that of FIG. 1, and the description of the same symbols and signal names as those used in the first embodiment will be omitted. The original G signal and R signal change in opposite directions,
FIG. 3 shows a signal waveform when the R signal is used to remove signal noise.

With respect to the original signal GS shown in FIG.
As shown in FIG. 3D, it is assumed that the third and fourth samples of the input signal GI are noise. It is also assumed that an input signal RI as shown in FIG.

First, the input signal G including such noise
When I is given to the low-pass filter 1 of FIG. 1, the low-pass filter 1 performs averaging of the signals two samples before and two samples later, and as a low-frequency component of the G signal as shown in FIG. GL is generated. On the other hand, when the input signal RI is given to the high-pass filter 1, so that the opposite characteristics of the low-pass filter 1 is obtained from the current signals, two samples before and two samples after the signal averaging signal subtracted signal, i.e. a high-frequency component of the R signal is the signal RH, as shown in FIG. 3 (3) is generated.

The high-frequency component processing circuit 3 examines the increase / decrease of the signal after two samples compared to the signal two samples before, and the G signal increases and the R signal decreases, or the G signal decreases and the R signal increases. In this case, a signal obtained by inverting the signal RH is output, and otherwise, the signal RH is output as it is. Thus, the high-frequency component processing circuit 3 generates a signal GH as shown in FIG. 3 (6) as a corrected high-frequency component of the G signal. Addition circuit 5
Adds the signal GH and the signal GL, and outputs a correction signal GC of the G signal as shown in FIG.

[0027] The selection circuit 5 with respect to the thus obtained signal GC
Selects the signal GC when the noise range designation signal NF shown in FIG. 3 (8) is 1, and selects the input signal GI when the signal NF is 0. As a result, as shown in FIG. 3 (9), a signal GO in which a noise portion is replaced is obtained as an output signal of the G signal. Here, the signal NF is 1 when the third and fourth samples are noise, and is 0 when the first, second, fifth, and sixth samples are not noise.

By the above operation, the original signals are R, G,
No correlation between B, even when the the G signal and the R signal is changed to the opposite direction, it is possible to reduce the noise.

Embodiment 3 Next, a noise removing circuit according to Embodiment 3 of the present invention will be described with reference to the drawings. FIG. 4 is a block diagram showing a configuration of the noise elimination circuit according to the present embodiment. About what is identical to the block and signal names of the symbols used in the first embodiment,
Description is omitted.

This noise elimination circuit includes a low-pass filter 1, a first high-pass filter 2R and a second high-pass filter 2B, and a high-pass filter 2R having characteristics opposite to those of the low-pass filter 1. A high-frequency component processing circuit 3 that selects and outputs either the output of 2B or 0, an addition circuit 4 that adds the output signal of the low-pass filter 1 and the output signal of the high-frequency component processing circuit 3, and And a selection circuit 5 for selecting and outputting any of the video signal and the output signal of the addition circuit 4 in accordance with the control signal.

In the noise removing circuit configured as described above, the operation when the original G signal and the B signal change in the same direction and the R signal changes in the opposite direction to the G signal will be described. . FIG. 5 and FIG. 6 are explanatory diagrams showing signal waveforms of various parts under such conditions. FIG.
FIG. 4 is an explanatory diagram of judgment conditions performed by the high-frequency component processing circuit 3.

With respect to the original signal GS shown in FIG.
As shown in FIG. 5 (4), it is assumed that the three samples and the fourth signal GI to th sample becomes noise is input. It is also assumed that the signal RI in FIG. 5B and the signal BI in FIG. 5C are input.

First, an input signal GI is used as a first video signal.
Is applied to the low-pass filter 1 in FIG. 4, the low-pass filter 1 performs averaging of the signals two samples before and two samples later, and obtains a low-frequency component of the G signal as shown in FIG. Generate a signal GL. Also, as the second video signal, the input signal RI as shown in FIG.
, The high-pass filter 2R subtracts a signal obtained by averaging the signal two samples before and two samples after the current signal, in order to perform the reverse characteristic processing of the low-pass filter 1, FIG. 6 (5) as the signal of the first high frequency component
Is generated. When the input signal BI of FIG. 6 (2) is given as the third video signal by the same processing as the R signal, the high-pass filter 2B outputs the signal BH shown in FIG. 6 (4) as the second high-frequency component. Generate

Then, the high-frequency component processing circuit 3 determines whether the signal after two samples has increased or decreased compared to the signal two samples before the sample of interest, and according to the table shown in FIG.
H, any one of the signals BH and 0 is output. Here, 3
In the sample and the fourth sample, the G signal increases, the B signal also increases, and the R signal decreases. Therefore, the change in the G signal and the change in the R signal are different, and the change in the G signal and the change in the B signal match. Will be. In this case, the high-frequency component processing circuit 3 selects the signal BH and uses it as a corrected high-frequency component of the G signal in FIG.
Is generated as shown in FIG. And the addition circuit 4
Adds the signal GH and the signal GL, and outputs a signal GC shown in FIG. 5 (7) as a correction signal of the G signal.

In response to the signal GC thus obtained, the selection circuit 5 sets the noise range designation signal NF shown in FIG.
, The signal GC is selected, and when the signal NF is 0, the input signal GI is selected. As a result, as the output signal of the G signal, a signal GO in which a noise portion is replaced is obtained as shown in FIG. Here, the signal NF is a signal which is 1 when the third and fourth samples are noise, and is 0 when the first, second, fifth and sixth samples are not noise.

By the above operation, the original signals are R, G,
Even when there is no correlation between B and the G signal and the R signal change in opposite directions, noise can be reduced. In the above description, it is assumed that noise is included at a specific position of the first video signal, that is, the G signal. However, the signal including noise may be an R signal or a B signal.

[0037]

As described above, according to the noise elimination circuit of the present invention, even if the signal containing noise and the other signal are changing in the same direction or in the opposite direction. , Noise can be made less noticeable.

[Brief description of the drawings]

1 is a configuration diagram of the noise removal circuit in the first and second embodiments of the present invention.

FIG. 2 is a signal waveform diagram illustrating an operation of the noise removal circuit according to the first embodiment.

FIG. 3 is a signal waveform diagram illustrating an operation of the noise removal circuit according to the second embodiment.

FIG. 4 is a configuration diagram of a noise removal circuit according to a third embodiment of the present invention.

Is a [5] signal waveform diagram representing an operation of the noise removal circuit of the third embodiment (part 1).

FIG. 6 is a signal waveform diagram (part 2) illustrating an operation of the noise removal circuit according to the third embodiment.

FIG. 7 is an explanatory diagram illustrating processing of a high-frequency component processing circuit in the noise removal circuit according to the third embodiment;

8 is a block diagram of a noise reduction circuit of the conventional example.

FIG. 9 is a signal waveform diagram showing an operation of the conventional noise elimination circuit.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Low-pass filter 2, 2B, 2R High-pass filter 3 High-pass component processing circuit 4 Addition circuit 5 Selection circuit

Claims (3)

[Claims] 1. A noise elimination circuit that suppresses a noise component when an inherent noise of the image sensor exists in an image signal including first and second video signals output simultaneously from the image sensor. A low-pass filter that extracts a low-frequency component from the first video signal including the noise component; a high-pass filter that extracts a high-frequency component from the second video signal that does not include the noise component; wherein when the first polarity of the change amount of the second video signal of the variation of the video signal is the same is output as the high-frequency component, wherein the polarity of the change amount of the first video signal the 2
High-frequency component processing means for attenuating and outputting the high-frequency component of the high-pass filter when the polarity of the change amount of the video signal is different, and outputting the low-frequency component and the output signal from the high-frequency component processing means. Adding means for adding, and selecting means for selecting an output signal from the adding means when the first video signal is noise, and selecting the first video signal when the first video signal is not noise, according to a signal indicating a noise portion. And a noise elimination circuit. 2. A noise elimination circuit that suppresses a noise component when an inherent noise of the image sensor exists in an image signal including first and second video signals output simultaneously from the image sensor. , a low pass filter for extracting a low frequency component from said first video signal including the noise component, and a high-pass filter for extracting a high frequency component from the second video signal not including the noise component, the When the polarity of the change amount of the first video signal is the same as the polarity of the change amount of the second video signal, the high-frequency component is output as it is, and the polarity of the change amount of the first video signal and the second
High-frequency component processing means for inverting and outputting the high-frequency component when the polarity of the change amount of the video signal is different; andaddition means for adding the low-frequency component and the output signal from the high-frequency component processing means. Selecting means for selecting an output signal from the adding means when the first video signal is noise according to a signal indicating a noise portion, and selecting the first video signal when the first video signal is not noise. A noise removal circuit characterized by the above-mentioned. 3. A noise elimination circuit that suppresses the noise component when an image signal including first, second, and third video signals output simultaneously from the image sensor includes noise inherent in the image sensor. A low-pass filter for extracting a low-frequency component from the first video signal containing the noise component, and a first low-pass filter from the second video signal containing no noise component.
A first high-pass filter that extracts a high-frequency component of the first video signal; a second high-pass filter that extracts a second high-frequency component of the third video signal that does not include the noise component; When the polarity of the change amount of the video signal is the same as the polarity of the change amount of the second video signal, the first high-frequency component is selected, and the polarity of the change amount of the first video signal and the second If the polarity of the change amount of the video signal is different and the polarity of the change amount of the first video signal is the same as the polarity of the change amount of the third video signal, the second high-frequency component is selected. The polarity of the change amount of the first video signal is different from the polarity of the change amount of the second video signal, and the polarity of the change amount of the first video signal is different from the polarity of the change amount of the third video signal. In this case, a signal obtained by attenuating one of the first high-frequency component and the second high-frequency component is selected and output. To the high-frequency component processing means, the addition means for adding the output signal from the high frequency component processing means and the low frequency component, when the following signals indicating the noise portion first video signal is a noise above Selecting means for selecting an output signal from the adding means and selecting the first video signal when the signal is not noise.