JP2006120891A - Evaluation method and device of solid state imaging element - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an evaluation method and evaluation device capable of automatically accurately and rapidly judging weak contrast defects, the so-called stain and unevenness, causing serious problems in images in a solid state imaging element. <P>SOLUTION: Image data from the solid state imaging element is used to provide, as one set, a binary threshold value that takes a difference of thickness as an index, and an area value that takes a defect feature amount as an index. At least two sets of these values are combined to discriminate a point-shaped defect having an equivalent thickness difference to a defect object and high frequency band noise from stain and unevenness, for judgement of acceptance of the solid state imaging element. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an evaluation method and an evaluation apparatus for a solid-state imaging device that inspects and determines the electrical characteristics of the solid-state imaging device.

  In recent years, solid-state imaging devices have made remarkable progress in the consumer and industrial fields, and in particular, almost 100% solid-state imaging devices are used in digital cameras and video cameras.

  The solid-state imaging device usually has provisions for image quality in the electrical characteristics, and it is necessary to accurately and quickly inspect and select products that satisfy these provisions.

  As a means for inspecting and selecting the electrical characteristics of the solid-state imaging device, it is common to first select automatically in a wafer state by a tester and then select by a tester or a mounting evaluation apparatus after assembly.

  Accordingly, a defect inspection method for a solid-state imaging device using a tester will be described based on the inspection flowchart of FIG. First, the solid-state image sensor is operated (F41), and image data that is a signal from the solid-state image sensor is stored for each pixel in an image storage device inside or outside the tester (F42). Binarization is performed with a threshold value TH1 obtained by multiplying the average value AVG of the entire stored image data by a predetermined ratio (F43), a defect area is calculated (F44), and the defect area is determined as an area standard S1. If it is smaller (No in F45), it branches to a non-defective product (F46). If it is larger than the area standard S1 (Yes in F45), it is determined as a defective product (F47).

  For example, the threshold value TH1 is set to about ± 3% of the average value AVG. In this case, it is determined whether or not the average value AVG is in the range of 97% to 103%. Here, when it is smaller than 100%, it is detected as dark side noise, and when it is larger than 100%, it is detected as bright side noise.

Further, as shown in the image of FIG. 3, there are defects in which a change in signal level appears locally called a scratch and a defect in which a signal level change appears in a wide area called a spot. Here, scratches can be corrected by interpolation using signal processing. However, since it is difficult to take measures against stains and unevenness, scratches are treated as non-defective products, and stains and unevenness are defective products. Need to be sorted as.
JP 2004-15712 A

  However, according to the above-described conventional evaluation method, since the defect is determined by one determination using the threshold value calculated from the average value of the image data, the contrast is clear with respect to the average value. Defect determination is possible. However, it is difficult to selectively determine defects with low contrast, which are problems in images, so-called spots and unevenness. That is, if the determination condition is relaxed, a lot of base noise is detected, and as a result, the detection accuracy deteriorates. Further, if the judgment conditions are strict, it may not be detected, and it is very difficult to judge scratches, spots, and unevenness.

  For this reason, the spot / unevenness inspection has to be performed by inputting the image data from the solid-state imaging device to a monitor television and visually observing it, and it has been difficult to perform accurate and rapid sorting.

  SUMMARY OF THE INVENTION The present invention solves the above-described problems, and an object of the present invention is to provide an evaluation method and an evaluation apparatus for a solid-state imaging device capable of determining defects with low contrast, so-called spots / unevenness.

  In order to solve the above problems, a method for evaluating a solid-state imaging device according to the present invention includes a step of storing a pixel signal level output from each pixel of a solid-state imaging device having a plurality of pixels arranged two-dimensionally; Selecting a first defective pixel out of the first signal level range from the plurality of pixels, and determining the first defective area from the number of pixels adjacent to each other on the array in the first defective pixel. A step of calculating, a step of selecting a second defective pixel in which the first defective area is equal to or greater than a first area range among the first defective pixels, and a second signal among the plurality of pixels. Selecting a third defective pixel that deviates from the level range, calculating a second defective area from the number of pixels adjacent to each other on the array in the third defective pixel, and the third defective pixel Among the pixels, the second defect area is Is 2 areal extent or more, and characterized in that it comprises the step of selecting a fourth defective pixel including the second defective pixel.

  Here, the second area range is larger than the first area range.

  In order to solve the above problems, the solid-state imaging device evaluation apparatus according to the present invention stores an image memory that stores a pixel signal level output from each pixel of a solid-state imaging device having a plurality of pixels arranged two-dimensionally. And a first binarization unit that selects a first defective pixel that deviates from a first signal level range among the plurality of pixels, and the first defective pixel is adjacent to each other on the array. First defect area calculating means for calculating a first defect area from the number of pixels to be processed, and among the first defective pixels, a second defective pixel whose first defective area is not less than a first area range is selected. A first area judging means for selecting; a second binarizing means for selecting a third defective pixel that deviates from a second signal level range among the plurality of pixels; and the third defective pixel. , From the number of pixels adjacent to each other on the array, A second defect area calculating means for calculating an area; and a fourth defect area of the third defect pixel, wherein the second defect area is not less than a second area range and includes the second defect pixel. And a second area determining means for selecting defective pixels.

  Here, the second area range is larger than the first area range.

  According to the present invention, at least two stages of determination that combines the density difference and the defect area are performed, so that a defect having low contrast, which is a problem in an image, that is, a so-called spot / unevenness is automatically determined by a tester without visual inspection. It is possible to determine. In order to determine whether or not the solid-state image sensor is acceptable by combining the density difference and the area as an index, a threshold value is calculated from the data average value as in the past, and this threshold value is compared with the above data average value. Unlike the method, it is possible to accurately and quickly determine a defect having low contrast, which is equivalent to a point-like scratch or a high-frequency base noise, and has a low contrast, which is a problem in an image, so-called spots / unevenness.

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

  FIG. 1 is an inspection flowchart for explaining a solid-state image sensor evaluation method according to an embodiment of the present invention, and FIG. 2 is a block diagram of a solid-state image sensor evaluation apparatus.

  FIG. 3 shows the results obtained when the inspection flowchart of FIG. 1 is executed, using an example in which spots and scratches are relatively compared.

  FIG. 4 shows an example for explaining the concept of adjusting the threshold value and the area standard when the inspection flowchart of FIG. 1 is executed.

  In the following embodiments, noise is output in the positive direction, but the same applies to the case where noise is output in the negative direction.

  As shown in FIG. 2, the evaluation apparatus according to the present embodiment captures a solid-state imaging device 1 having a plurality of pixels arranged two-dimensionally and photoelectrically converting an input image, and a pixel signal output from each pixel. The first storage unit 2 for detecting the first defective pixel by binarizing each pixel signal stored in the image storage unit 2 with the first threshold value TH1. First defect area calculation means 4 for calculating the first defect area from the number of pixels adjacent to each other on the pixel array in the first defective pixel, and the first defective area of the first defective pixel is the first defective area. A first area judging means 5 for selecting a second defective pixel having an area range equal to or larger than the above-mentioned area range; and a third defect obtained by binarizing each pixel signal stored in the image storage means 2 with a second threshold value THn. The second binarizing means 6 for detecting pixels and the third defective pixel are mutually connected on the pixel array. Second defect area calculating means 7 for calculating the second defect area from the number of pixels adjacent to the second defect area, and the second defect area of the third defective pixels is equal to or greater than the second area range, and And a second area determining means 8 for selecting a fourth defective pixel including the defective pixel.

  The evaluation method of this embodiment is demonstrated using FIG. The solid-state imaging device has a plurality of pixels arranged two-dimensionally, and a pixel signal output from each pixel is taken into a tester. The entire pixel signal is captured by the tester as image data (F1), and this image data is stored in an image storage device inside or outside the tester (F2). Thereafter, binarization (F3) is performed using the first threshold value TH1, and the defect area calculation is performed by detecting the size of the area of the adjacent portion exceeding the threshold value TH1 (F4). Here, the first threshold value TH1 is generally determined by multiplying the average value of the entire image data stored in the image storage device by a predetermined ratio. For example, the threshold value is set to about + 3% of the average value AVG. In this case, 103% or more of the average value AVG is determined as noise. Note that the first threshold value TH1 may be a fixed value.

  Then, the defect area of the adjacent portion exceeding the first threshold value TH1 is compared with the area standard S1 (F5). If the area is smaller than S1, the defect area is determined to be non-defective (F9). As screened. In the present embodiment, both scratches and spots are detected in a mixed state.

  Further, binarization by the second threshold value THn (F6) is performed, and the defect area calculation is performed by detecting the size of the area of the adjacent portion exceeding the threshold value THn (F7). The detected defect area is compared with the area standard Sn (F8), and if the area is smaller than Sn, it is determined as a non-defective product (F9), which includes defective pixels selected as spot candidates and the area is larger than Sn. A thing is determined to be a defective product (F10), and a pass / fail determination is performed. The area standard Sn is preferably determined as Sn> S1 based on the area standard S1.

  The result of executing the inspection flowchart of FIG. 1 will be described by comparison using the spot image 11 and the scratch image 21 shown in FIG. The density distribution 12 of the stain and the density distribution 22 of the flaws in which the density of the defect is represented by a three-dimensional graph are examples in which the density of the defect is represented by the z coordinate with the image space as the x and y coordinates. In order to extract a spot defect, binarization is performed with a threshold value TH1. Regarding the spot density distribution 12 and the flaw density distribution 22, a graph in which a pixel equal to or higher than the threshold TH1 is 1 and a pixel less than the threshold TH1 is 0 is a threshold binarization graph 13 and a threshold 2 It is a value graph 23. Further, a graph in which the density distribution 12 of the stain and the density distribution 22 of the scratch are set to 1 for pixels equal to or higher than the threshold THn and 0 for pixels lower than the threshold THn is a binarized graph 14 at the threshold THn. And a binarized graph 24.

  Next, a specific threshold setting method will be described with reference to FIG. FIG. 4A is a graph of the scratch image 11 and the scratch image 21 when the scratch and the stain are viewed in the x-direction cross section. Here, the coordinate 35 is on the horizontal axis, and the contrast 36 is on the vertical axis. The difference in the density of the scratch is indicated by the scratch 31, and the difference in the density of the spot is indicated by the spot 32. In order to determine a region having noise that exceeds a predetermined value in the measured light and shade distribution, the threshold value TH1 is first adjusted and set so that it can be binarized and detected at a higher light and shade distribution level. .

  FIG. 4B is a graph showing position coordinates detected as a result of binarization with the threshold value TH1. According to the result detected by binarization, the scratch 31 has a defect area value of 1 at the coordinate = 5, and the spot has a defect area value of 1 at the coordinate = 17.

  According to the result detected by binarization, the scratch 31 has a defect area value of 1 at the coordinate = 5, and the spot has a defect area value of 1 at the coordinate = 17.

  FIG. 4C is a graph showing position coordinates detected as a result of binarization with the threshold value THn. The scratch has a defect area value of 1 at coordinate = 5. As a result of detection of spots at coordinates = 15 to 19, the value of the defect area is 5.

  In the case of a flaw, since the spatial change of the measured value is steep, there is almost no difference in the defect area between the threshold TH1 and THn. Since such a change is gradual, there is a feature that a difference in defect area is greatly increased between the threshold TH1 and THn.

  Therefore, paying attention to the defect area difference obtained here, the value of the area standard Sn is set (F8). For example, if Sn is set to a value Sn = 3 larger than S1 (= 1), scratches and spots can be separated.

  As described above, by using the two-stage thresholds (TH1, THn) and the area standard (S1, Sn) for sorting, scratches and spots are separated and efficiently classified into good products and defective products. be able to.

  In the above embodiment, the stain is determined in two stages (TH1, S1) and (THn, Sn). However, the stain is determined by repeating the same processing a plurality of times. It does not matter.

  As described above, according to the present embodiment, it is possible to automatically determine a defect having a low contrast, that is, a so-called spot / unevenness, by combining the density difference and the area as an index in combination.

  The present invention is useful as a technique for automating sensory tests depending on human subjectivity.

Inspection flow chart for explaining a solid-state image sensor evaluation method in an embodiment of the present invention 1 is a block diagram of an evaluation apparatus for a solid-state image sensor according to an embodiment of the present invention. Block diagram for explaining a conventional method for evaluating a solid-state imaging device Explanatory diagram showing how to set the threshold Inspection flowchart for explaining a conventional method for evaluating a solid-state imaging device

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Solid-state image sensor 2 Image memory | storage means 3 1st binarization means 4 1st defect area calculation means 5 1st area judgment means 6 2nd binarization means 7 2nd defect area calculation means 8 2nd Area determination means 11 Spot image 12 Spot density distribution 13 Spot binarization graph (threshold value TH1)
14 Binarization graph (threshold value THn)
15 threshold 1
16 threshold n
21 Scratch image 22 Scratch density distribution 23 Scratch binarization graph (threshold value TH1)
24 Scratch binarization graph (threshold value THn)
31 Scratches 32 Spots 33 Threshold TH1
34 Threshold TH2
35 Coordinates 36 Gradation distribution 37 Binary count F1 Solid-state imaging device F2 Image storage device F3 Binary F4 by threshold TH1 Defect area calculation means F5 Area comparison means F6 Binary conversion by threshold THn Defect area calculation means F8 Area comparison means F9 Non-defective product judgment processing means F10 Defective product judgment processing means F41 Solid-state imaging device F42 Image storage device F43 Binary binarization by threshold value F44 Defect area calculation F45 Area standard F45 Non-defective product F46 Defective product

Claims (4)

Storing a pixel signal level output from each pixel of a solid-state imaging device having a plurality of pixels arranged two-dimensionally;
Selecting a first defective pixel out of the first signal level range from the plurality of pixels;
In the first defective pixel, calculating a first defective area from the number of pixels adjacent to each other on the array;
Selecting a second defective pixel in which the first defective area is equal to or greater than a first area range among the first defective pixels;
Selecting a third defective pixel out of the second signal level range from the plurality of pixels;
Calculating a second defect area from the number of pixels adjacent to each other on the array in the third defective pixel;
Selecting the fourth defective pixel including the second defective pixel, the second defective area being equal to or larger than the second area range among the third defective pixels. For evaluating a solid-state imaging device.   The method for evaluating a solid-state imaging device according to claim 1, wherein the second area range is larger than the first area range. Image storage means for storing a pixel signal level output from each pixel of a solid-state imaging device having a plurality of pixels arranged two-dimensionally;
First binarization means for selecting a first defective pixel out of the first signal level range from the plurality of pixels;
First defect area calculating means for calculating a first defect area from the number of pixels adjacent to each other on the array in the first defective pixel;
A first area determining means for selecting a second defective pixel having a first defective area equal to or larger than a first area range among the first defective pixels;
A second binarizing means for selecting a third defective pixel out of the second signal level range from the plurality of pixels;
Second defect area calculation means for calculating a second defect area from the number of pixels adjacent to each other on the array in the third defective pixel;
Second area determination means for selecting a fourth defective pixel including the second defective pixel, wherein the second defective area is equal to or larger than a second area range among the third defective pixels; An evaluation apparatus for a solid-state imaging device, comprising:   The solid-state imaging device evaluation apparatus according to claim 3, wherein the second area range is larger than the first area range.

Images