JP2850508B2 - Method for detecting line defect in solid-state image sensor - Google Patents

Description

Description: BACKGROUND OF THE INVENTION The present invention relates to a defect detection method for a solid-state imaging device for selecting whether a manufactured solid-state imaging device is a good product or a defective product. About the method.

[Summary of the Invention]

The present invention relates to a method for detecting a line defect of a solid-state imaging device in which a plurality of pixels are arranged in a matrix, wherein a difference between data subjected to vertical addition processing and data surrounding the data is determined for each vertical line, The detection accuracy of the line defect is improved by detecting the line defect from the comparison with the value.

[Conventional technology]

2. Description of the Related Art A solid-state imaging device such as a CCD imager is a device that is manufactured using a semiconductor manufacturing technology for performing fine processing on a semiconductor substrate and generates an electric signal according to incident light. In such a solid-state imaging device, a line defect along a vertical line may occur among a plurality of pixels arranged in a matrix due to various causes in a semiconductor manufacturing process. Therefore, in order to ship as a product, it is necessary to detect whether or not there is a line defect. Conventionally, line defects have been detected using an automatic defect detection device according to the following procedure. .

First, data from each pixel of the manufactured CCD imager is read, and the data (D image) is stored in, for example, a memory. This read data (D image)
If, for example, there are also line defects and point defects,
As schematically shown in FIG. 5A, these line defects d _L
DOO point defect d _P is to appear in the data (D image).

Next, the data (D image) is divided in the V direction (vertical direction) into sections of, for example, 5 pixels (5 bits), and 5-bit median data (DM image) is obtained from the median located at the center of the section. . At this time, the point defect d _P disappear, only a line defect d _L is a defect along the V direction will remain as shown in the FIG. 5 (b).

Next, 5-bit median data (DM image) is added in the V direction and averaged to obtain data (L) for one horizontal line as shown in FIG. 5 (c). Then, the line defect d _L remains as a data portion having a small value even when the addition and averaging is performed, so that the line defect d _L is directly detected as line defect data.

[Problems to be solved by the invention]

However, in the case of the above-described line defect detection method, H
When there is shading in the direction (horizontal direction), there arises a problem that the minimum value due to the shading is erroneously detected.

That is, as shown in FIG. 6A, when a line defect d _L exists in the 5-bit median data (DM image) and shading occurs in the H direction, a valid signal is generated in that portion. Can not be obtained.
Accordingly, when data (L) for one horizontal line is obtained by directly averaging data having shading in the V direction, line defect data is obtained as shown in FIG. Changes in the overall data according to the shading are also detected, for example, so that the end of the data (L) for one horizontal line has a minimum value (D _min ).
As a result, in the automatic defect detection device, the minimum value (D _min )
Was determined to be a line defect portion, which hindered detection of a true line defect.

In view of the above technical problem, an object of the present invention is to provide a method of detecting a line defect of a solid-state imaging device that accurately detects a line defect even when shading occurs.

[Means for solving the problem]

First, the solid-state imaging device used in the line defect detection method of the present invention includes a plurality of pixels arranged in a matrix. Here, the plurality of pixels may have a structure having the same filter alternately for each vertical column.

The method of detecting a line defect of a solid-state imaging device according to the present invention reads out data of a plurality of pixels arranged in a matrix, and adds or averages data of pixels in a pixel row in a vertical direction to obtain a representative value in the vertical direction. The representative value is taken as data of one horizontal line arranged in the horizontal direction, and the filter configuration of the pixel row in the vertical direction is the same, and the difference between the representative value and the representative value arranged one after the other in the horizontal direction is obtained. The sum or the average of each difference is obtained for each representative value constituting the data of the one horizontal line, and the sum of the differences or the average of the differences obtained for each representative value is compared with a predetermined value.

Finally, in the method for detecting a line defect of a solid-state imaging device according to the present invention, the line defect is detected by comparison with a predetermined value.

[Action]

In the defect detection method for the solid-state imaging device according to the present invention, the filter configuration of the pixel rows in the vertical direction is the same, and the sum of the respective differences from the representative values arranged in the horizontal direction is equal to or the average of the respective differences. Is obtained for each representative value constituting the data of one horizontal line. Then, the sum of the differences or the average of the differences obtained for each representative value is compared with a predetermined value. To detect.
Therefore, in a case where the reference state changes gradually, such as when shading occurs, the change in the reference state can be offset.

〔Example〕

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

The present embodiment is a method of detecting a line defect of a CCD imager that performs a difference using pixels of about 2 bits.

First, an example of a CCD imager used in the present embodiment will be briefly described. The CCD imager has a structure as shown in FIG. The CCD imager includes a plurality of light receiving sections 21, a plurality of vertical register sections 22, a horizontal register section 23, and an output section 24.
And formed on a silicon substrate by using a fine processing technique. The light receiving sections 21 are arranged in a matrix corresponding to each pixel. Each of these light receiving sections 21 is composed of a photo sensor, and generates signal charges according to incident light. The vertical register section 22 is formed for each vertical column of the light receiving section 21 and transfers signal charges in the vertical direction. The horizontal register unit 23 is a register that is electrically connected to each of the vertical register units 22 and transfers the signal charges from each of the vertical register units 22 for each horizontal line. And
At the output unit 24, the signal charges are converted into image signals and read.

Although not shown, this CCD imager is for a color signal and has three R (red), G (green), and B (blue) signals.
Two filters are formed on the chip. And
In the arrangement of the filters, a row composed of only G (green) filters and a row in which R (red) and B (blue) filters are alternately formed for each pixel are alternately arranged in the horizontal direction. The array is as follows. That is, the filter arrangement of the vertical line shifted by 2 bits in the horizontal direction is the same.

In the CCD imager having such a structure, a line defect that makes it impossible to effectively read pixels of one vertical line may occur due to its manufacture. In the line defect detection method according to the present embodiment, the following procedure shown in FIG.
A line defect corresponding to the one vertical line is detected.

First, a CCD imager to be measured is set in the automatic defect detection device. After the setting, a signal for driving the CCD imager is given to the CCD imager to be measured, and data (D) for all pixels is read. This data (D) is temporarily stored in the memory of the detection device (ST
1).

Next, the stored data (D) is divided into sections of 5 bits each in the V direction (vertical direction), and 5-bit median data, that is, a median (DM) located at the center of the section is extracted ( ST2). By taking out the median, most point defects will disappear.

Next, the central value (DM) of every 5 bits is averaged in the V direction to calculate a representative value (L) of data for each vertical line.
The representative value of the data for each vertical line constitutes data for one horizontal line. Here, when the one vertical line has a line defect, the representative value (L) is, for example, data of one horizontal line distributed on the time axis as shown in FIG. If the line defect is determined simply with the minimum value as it is in the prior art, a problem occurs when judging occurs as described above. In this embodiment, in addition to the conventional processing, the following difference is obtained. Since the process (ST4) is performed on one horizontal line segment data, a reliable detection of a line defect is realized.

Explaining the difference processing, the CCD imager measured in this embodiment has the same filter configuration when the RGB arrangement of the filter is shifted by 2 bits (2 vertical lines) as described above. Therefore, the data of a pixel that should have a difference with respect to the data of a certain vertical line is data that is horizontally shifted by about 2 bits. FIG. 3 schematically illustrates this difference processing. In the figure, the memory unit 10 is arranged so that each representative value is horizontally shifted from one representative value L of one horizontal line data arranged in the horizontal direction. And the memory unit 12 stores data Lr which is shifted from the representative value L by two bits in the horizontal direction to the rear side. This will be described with reference to the time axis. As shown by the waveforms in FIG. 3, the data (L
1) is data shifted to the left by 2 bits with respect to the memory unit 10, that is, data advanced by 2 bits in time,
The data (Lr) stored in the memory unit 12 is data shifted to the right by 2 bits with respect to the memory unit 10, that is, data delayed by 2 bits in time, and corresponds to a line defect in the representative value L. The value (D ₀ ) obtained in the memory unit 11 is a value shifted to the left by 2 bits, that is, at a time earlier by 2 bits.
Becomes D _1, position i.e. two bits shifted to the right by 2 bits in the memory unit 12 only just late would be the value D ₂ reversed. Therefore, if the data at a certain time (t ₀ ) is simultaneously extracted from the three memory units 10 to 12, the data stored in the memory unit 10 is two bits earlier than the data stored in the memory unit 10. The unit 12 processes data that is two bits later in time than the data stored in the memory unit 10.

Here, when the difference processing is expressed by an equation, in the present embodiment, processing of Lx = [(L-Ll) + (L-Lr)] / 2 is performed. That is, the data Ll after 2 bits
, And the difference between the data Lr two bits before, and averaging. In FIG. 3, the subtraction processing units 13 and 14
Each subtraction process in the above equation is performed, and the addition processing unit 15
The result of each subtraction process is added. Then, the data from the addition processing unit 15 is standardized (averaged) and stored in the memory unit 16.
Stores the processing result Lx. By taking the difference of each data and averaging, the change from the reference state is canceled. Therefore, when there is a line defect, a value D _L corresponding to the line defect can be obtained without influence of shading. As a result, in the automatic defect detecting device can detect a line defect in comparison with the defect detection level L _min, for each processing result Lx, respectively to compare the defect detection level L _min (ST5), after all, All processing results Lx are defect detection levels L
is greater than _min is determined as "line without disqualifying" Conversely, if even one processing result Lx is less than the defect detection level L _min will be determined that the "line with defect".

According to the detection method of this embodiment, the line defect of the CCD imager can be detected without the influence of shading.
The accuracy of determining a line defect is greatly improved.

In the above-described embodiment, the difference processing is performed by shifting back and forth by 2 bits. However, the method of detecting a line defect according to the present invention is not limited to this, and various types of surrounding data may be adjusted according to the configuration of the filter. It is also possible to take the difference with.

〔The invention's effect〕

As described above, the method for detecting a line defect of a solid-state imaging device according to the present invention employs the same filter configuration for the vertical pixel row, and the sum of the respective differences from the representative values arranged one after another in the horizontal direction. Alternatively, an average of the differences is obtained for each representative value constituting the data of one horizontal line. Then, the sum of the differences or the average of the differences obtained for each representative value is compared with a predetermined value. To detect. As a result, in the method for detecting a line defect of a solid-state imaging device according to the present invention, the deviation from the reference state is canceled, and the line defect can be reliably detected even when shading occurs. Therefore, by applying the present invention to a solid-state imaging device manufacturing line, it is possible to efficiently provide a solid-state imaging device having no line defect.

[Brief description of the drawings]

FIG. 1 is a flowchart showing an example of an algorithm according to a method for detecting a line defect of a solid-state imaging device according to the present invention, FIG. 2 is a waveform diagram showing an example of data having a line defect, and FIG. FIG. 4 is a schematic diagram for explaining an example of processing, FIG. 4 is a block diagram of a general CCD imager, and FIGS.
(C) is a diagram for explaining a conventional method for detecting a line defect of a solid-state imaging device, and (a) and (b) of FIG. 6 are diagrams for explaining a problem at the time of shading.

Claims (2)

(57) [Claims] 1. A method for detecting a line defect of a solid-state imaging device, comprising a plurality of pixels arranged in a matrix and reading data for each pixel, comprising the steps of: Is read, and the data of the pixels in the vertical pixel row are added or averaged to obtain a representative value in the vertical direction, and the representative value is taken as data of one horizontal line arranged in the horizontal direction, and the pixel row in the vertical direction is obtained. The sum of the respective differences or the average of the respective differences from the representative values having the same filter configuration and arranged in the horizontal direction is calculated for each representative value constituting the data of one horizontal line. Detecting a line defect of the solid-state imaging device by comparing a sum of the differences obtained for each representative value or an average of the differences with a predetermined value to detect a line defect of the solid-state imaging device. . 2. The solid-state imaging device has a filter configuration in which pixel columns in the vertical direction are alternately the same for each column, and includes two representative values arranged two pixels in the horizontal direction. The sum of each difference or the average of each difference,
The method for detecting a line defect of a solid-state imaging device according to claim 1, wherein the value is obtained for each representative value constituting the data of one horizontal line.