Chip surface defect detection method and system

The chip grayscale image is obtained through X-ray detection, grayscale mapping enhancement and filtering smoothing processing are performed, and the diffusion deviation value is calculated to judge the chip surface defects, which solves the problem of insufficient detection in the prior art and improves the accuracy of chip surface defect detection. Sex and effect.

CN118735913BActive Publication Date: 2024-11-15JIANGSU ETERN
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411205963.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-30
Publication Date
2024-11-15
Estimated Expiration
2044-08-30

AI Technical Summary

Technical Problem

Existing chip surface defect detection methods, especially X-ray photography detection, cannot accurately locate defect locations and low data processing accuracy, resulting in inaccurate detection and affecting chip performance and reliability.

Method used

X-ray detection is used to obtain the chip grayscale image, and the grayscale mapping enhancement, filtering smoothing processing and diffusion deviation calculation are used to determine whether the pixel points are abnormal and improve detection accuracy.

Benefits of technology

Through mapping equalization and filtering smoothing processing, the detection error rate is reduced, detection accuracy and effect are improved, and chip surface defects can be more accurately judged.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118735913B_ABST
    Figure CN118735913B_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of chip detection, and specifically discloses a chip surface defect detection method and system, comprising the following steps: obtaining a chip grayscale image of a target area by X-ray detection, obtaining the point grayscale value of a pixel point in the chip grayscale image of the target area; performing grayscale mapping enhancement on the grayscale value of each pixel point to obtain a point mapping grayscale value; performing filtering processing on the point mapping grayscale value of each pixel point and obtaining a point smoothing pixel value; traversing the pixel points in the chip grayscale image of the target area based on the point smoothing pixel value of the pixel point after filtering processing, and performing diffusion deviation value calculation on the point smoothing pixel value of the pixel point; judging whether the pixel points of the chip grayscale image of the target area are abnormal, and obtaining a pixel abnormality signal, and performing enhancement equalization and filtering smoothing processing on the chip grayscale image obtained by X-rays, so that the calculation comparison is more accurate, the misjudgment rate of detection is reduced, and the detection effect and accuracy are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of chip detection, and in particular to a chip surface defect detection method and system. Background Art

[0002] Chip surface defect detection is a crucial link in electronic chip manufacturing. The presence of chip surface defects may cause chip performance degradation, affecting its service life and reliability. Existing manual inspection methods, including visual inspection and microscope inspection, can only detect some relatively obvious defects. For minor defects, there are problems of low detection accuracy and slow detection speed.

[0003] In view of the shortcomings of manual inspection, the method of using machine vision inspection has gained more favor. At the same time, with the development of artificial intelligence, which has been very popular in recent years, the method of integrating deep learning methods with machine vision will make the surface defect detection method more intelligent.

[0004] The existing method of detecting by X-ray photography cannot locate the exact position of defects during the defect detection process, and the data processing accuracy is low during the image processing process, which is prone to inaccurate detection and affects the results of chip detection. Summary of the invention

[0005] The object of the present invention is to provide a chip surface defect detection method and system to solve the above-mentioned background problems.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A chip surface defect detection method comprises the following steps:

[0008] Step 1: Obtain a grayscale image of the chip in the target area using X-ray detection, and then mark the pixel points in the grayscale image of the chip in the target area as i, where i is 1, 2, 3, etc.; at the same time, perform grayscale value determination based on the total value L of the standard grayscale level, and obtain the point grayscale value H of the pixel point in the grayscale image of the chip in the target area i ; For each pixel i, the gray value H i Perform grayscale mapping enhancement to obtain the point mapping grayscale value GX i ;

[0009] Step 2: Map the gray value GX to each pixel i i Perform filtering and obtain the point smoothing pixel value HP i ;

[0010] Step 3: Smooth pixel value HP based on the pixel i after filtering i, traverse the pixel points in the grayscale image of the target area chip, and smooth the pixel value HP of pixel point i i Diffusion deviation value CCp i calculate;

[0011] Step 4: Diffusion deviation value CCp of point smoothing pixel value based on pixel i i ; The diffusion deviation value CCp of the point smoothing pixel value of pixel i i Compare with the diffusion deviation threshold CCpy; determine whether the pixel points of the grayscale image of the chip in the target area are abnormal, and obtain the pixel abnormality signal.

[0012] As a further solution of the present invention: In step 1, the gray value H of each pixel i is i Perform grayscale mapping enhancement to obtain the point mapping grayscale value GX i , including the following steps:

[0013] Based on each pixel point i and its corresponding point grayscale value, the point grayscale value frequency statistics are performed; the point grayscale value frequency set {DH n , r n};

[0014] Among them, DH n is the gray value of the point in the chip image of the target area, n is 1, 2, 3...; and n≤i; DH n As n is arranged from small to large; r n is the gray value of the point DH n Number of occurrences;

[0015] Based on the gray value of the point and the gray value frequency of the gray value of the point, a cumulative pixel value set {DH n , L.J. n}; Among them, LJ n Refers to the grayscale frequency r from the minimum grayscale value to the current grayscale value n The sum of; specifically, ;

[0016] At this time, based on the pixel i to pixel gray level mapping, the total gray level value DL is defined, where: ;

[0017] Based on the definition of the total gray level value DL, Calculate the gray value DH of each pixel n Mapping updated point mapping gray value GX (DH n );

[0018] Then we get the gray value H of each pixel i i Grayscale mapping enhancement obtains the point mapping grayscale value GX i.

[0019] As a further solution of the present invention: In the step 2, the gray value GX is mapped to the point i The steps of filtering process include:

[0020] B01: Constructing the Gaussian kernel formula of kXk , where the standard deviation δ is a preset value;

[0021] B02: Pass Calculate the normalization factor of the Gaussian kernel formula to obtain the normalization factor value ω of the pixel value filtering;

[0022] B03: Based on the Gaussian kernel formula and the normalization factor value, traverse the point mapping gray value GX of each pixel in the gray image of the chip in the target area i , filter the point mapping gray value of each pixel i to obtain the point smoothing pixel value HP i .

[0023] As a further solution of the present invention: in B01, k is greater than or equal to 3.

[0024] As a further solution of the present invention: In B03, the point mapping gray value of each pixel point i is filtered to obtain the point smoothing pixel value HP i ; The method is:

[0025] Taking pixel i as the center point, obtain the pixel matrix of kXk;

[0026] pass Map the gray value GX to the pixel i i Perform filtering calculation to obtain the point smoothing pixel value HP i , where G(x, y) is the Gaussian kernel; GXgs i The gray value matrix of the point kXk centered at pixel i.

[0027] As a further solution of the present invention: when the point mapping grayscale values ​​of the pixel points are traversally filtered, if the pixel points in the target area chip grayscale image are located at the edge of the target area chip grayscale image, the point mapping grayscale value matrix will exceed the boundary of the pixel points in the target area chip grayscale image; at this time, the part exceeding the boundary is filled with zeros, that is, the pixel points in the point mapping grayscale value matrix that exceed the target area chip grayscale image are regarded as zero.

[0028] As a further solution of the present invention: in step 3, the step of calculating the diffusion deviation value of the point smoothing pixel value of the pixel point includes:

[0029] Step W1: Get the point smoothing pixel value HP of pixel i i, and at the same time obtain the point smoothed pixel values ​​of the m adjacent pixels of pixel i, and mark them as HP in turn (i,m) ;

[0030] Step W2: Pass Calculate the point smoothing pixel value HP of pixel i i The point smoothing pixel value HP with adjacent pixels (i,m) Deviation value DP (i,m) ;

[0031] Step W3: Calculate the standard deviation of the deviation between the point smoothed pixel value of pixel point i and the point smoothed pixel values ​​of m adjacent pixels to obtain the diffusion deviation value CCp of the point smoothed pixel value of pixel point i i .

[0032] As a further solution of the present invention: the diffusion deviation value CCp of the point smoothing pixel value of pixel i is i Compare with the diffusion bias threshold CCpy;

[0033] If the diffusion deviation value CCp of the point smoothing pixel value of pixel i i If it is greater than the diffusion deviation threshold CCpy, a pixel abnormal signal is generated;

[0034] If the diffusion deviation value CCp of the point smoothing pixel value of pixel i i Less than or equal to the diffusion deviation threshold CCpy, a normal pixel signal is generated.

[0035] As a further solution of the present invention: based on the abnormal signal of the pixel point, it is prompted that the chip surface at the pixel point in the grayscale image of the chip in the target area is abnormal.

[0036] As a further solution of the present invention: a chip surface defect detection system, comprising:

[0037] Data acquisition module: used to obtain the chip grayscale image of the target area by X-ray detection, and then mark the pixel points in the chip grayscale image of the target area as i, where i is 1, 2, 3, etc.; at the same time, grayscale value determination is performed based on the total value L of the standard grayscale level to obtain the point grayscale value H of the pixel point in the chip grayscale image of the target area i ; For each pixel i, the gray value H i Perform grayscale mapping enhancement to obtain the point mapping grayscale value GX i ;

[0038] Data processing module: used to map the gray value GX to each pixel i i Perform filtering and obtain the point smoothing pixel value HP i ;

[0039] Deviation calculation module: based on the point smoothing pixel value HP of pixel i after filtering i , traverse the pixel points in the grayscale image of the target area chip, and smooth the pixel value HP of pixel point i i Diffusion deviation value CCp i calculate;

[0040] Abnormal comparison module: Diffusion deviation value CCp of point smoothing pixel value based on pixel i i ; The diffusion deviation value CCp of the point smoothing pixel value of pixel i i Compare with the diffusion deviation threshold CCpy; determine whether the pixel points of the grayscale image of the chip in the target area are abnormal, and obtain the pixel abnormality signal.

[0041] Beneficial effects of the present invention:

[0042] In the present invention, the difference between the point smoothed pixel value on the pixel point processed by mapping equalization and filtering smoothing and the point smoothed pixel value of the adjacent pixel point is calculated to obtain a diffusion deviation value, and then by judging whether the diffusion deviation value is within the diffusion deviation threshold, it is judged that the transition change between the pixel point and the adjacent pixel point is a smooth and stable change, so as to judge whether there is an abnormality at the position of the pixel point on the chip surface, and by performing enhanced equalization and filtering smoothing processing on the chip grayscale image acquired by X-rays, the calculation comparison is made more accurate, the error rate of detection is reduced, and the detection effect and accuracy are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The present invention will be further described below in conjunction with the accompanying drawings.

[0044] Figure 1 It is a schematic flow chart of the method of the present invention;

[0045] Figure 2 It is a system block diagram of the present invention. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Embodiment 1

[0047] See also Figure 1 As shown, the present invention is a chip surface defect detection method, comprising the following steps:

[0048] Step 1: Obtain a grayscale image of the chip in the target area using X-ray detection, and then mark the pixel points in the grayscale image of the chip in the target area as i, where i is 1, 2, 3, etc.; at the same time, perform grayscale value determination based on the total value L of the standard grayscale level, and obtain the point grayscale value H of the pixel point in the grayscale image of the chip in the target area i ; For each pixel i, the gray value H i Perform grayscale mapping enhancement to obtain the point mapping grayscale value GX i ;

[0049] The specific steps include:

[0050] Based on each pixel point i and its corresponding point grayscale value, the point grayscale value frequency statistics are performed; the point grayscale value frequency set {DH n , r n};

[0051] Among them, DH n is the gray value of the point in the chip image of the target area, n is 1, 2, 3...; and n≤i; DH n As n is arranged from small to large; r n is the gray value of the point DH n Number of occurrences;

[0052] Based on the gray value of the point and the gray value frequency of the gray value of the point, a cumulative pixel value set {DH n , L.J. n}; Among them, LJ n Refers to the grayscale frequency r from the minimum grayscale value to the current grayscale value n The sum of; specifically, ;

[0053] At this time, based on the pixel i to pixel gray level mapping, the total gray level value DL is defined, where: ;

[0054] Based on the definition of the total gray level value DL, Calculate the gray value DH of each pixel n Mapping updated point mapping gray value GX (DH n );

[0055] Then we get the gray value H of each pixel i i Grayscale mapping enhancement obtains the point mapping grayscale value GX i ;

[0056] By using grayscale mapping to process the acquired chip grayscale image of the target area, the grayscale value of each pixel of the chip grayscale image of the target area can be effectively equalized, thereby enhancing the image contrast and achieving the purpose of enhancing the contrast;

[0057] Step 2: Map the gray value GX to each pixel i i Perform filtering and obtain the point smoothing pixel value HP i , the steps of filtering process include:

[0058] B01: Constructing the Gaussian kernel formula of kXk , wherein the standard deviation δ is a preset value; in some embodiments, the value of δ is set according to the uniformity of the distribution of the chip image, so as to make the target area chip grayscale image smoother; the Gaussian algorithm is used for filtering, so that the edge features of the collected target area chip grayscale image can be retained during the smoothing calculation of the image pixels of the target area chip grayscale image; wherein k is greater than or equal to 3;

[0059] For example, in some calculations, G(x, y) is set to a 3X3 coordinate point; x is 0, 1, 2; y is also 0, 1, 2; the 3X3 Gaussian kernel formula has (x, y) values ​​of 9 coordinate points from (0, 0) to (2, 2); and in the calculation process, the center point (1, 1) of the matrix is ​​taken as the center;

[0060] When used to process chip grayscale images, the 3X3 Gaussian kernel G(x, y) can be expressed as:

[0061]

[0062] In some embodiments, for example, the matrix in a Gaussian kernel with a standard deviation δ of 1 is: ;

[0063] B02: Pass again Calculate the normalization factor of the Gaussian kernel formula to obtain the normalization factor value ω of the pixel value filtering;

[0064] For example, in the calculation of the normalization factor of a 3X3 Gaussian kernel with a standard deviation of δ of 1, ω is 0.93;

[0065] B03: Based on the Gaussian kernel formula and the normalization factor value, traverse the point mapping gray value GX of each pixel in the gray image of the chip in the target area i , filter the point mapping gray value of each pixel i to obtain the point smoothing pixel value HP i ;

[0066] Specifically: Take pixel point i as the center point, obtain the pixel point matrix kXk, and obtain the point mapping gray value matrix GXgs kXk with pixel point i as the midpoint i ;

[0067] Also taking the 3×3 example, the step of filtering the point mapping grayscale value of each pixel point i to obtain the point smoothing pixel value includes:

[0068] For example, taking pixel i as the center point, obtain a 3X3 pixel matrix, and at the same time obtain 9 pixel point mapping grayscale values ​​to construct a point mapping grayscale value matrix of pixel i. ;

[0069] pass Map the gray value GX to the pixel i i Perform filtering calculation to obtain the point smoothing pixel value HP i ; Where G(x, y) is the Gaussian kernel; GXgs i Map the gray value matrix to the kXk point centered at pixel i;

[0070] It should be noted that when performing traversal filtering on the point mapping grayscale values ​​of the pixel points in the target area chip grayscale image, if the pixel points in the target area chip grayscale image are located at the edge of the target area chip grayscale image, the point mapping grayscale value matrix will exceed the boundary of the pixel points in the target area chip grayscale image; at this time, zero filling is performed on the portion exceeding the boundary, that is, the pixel points in the point mapping grayscale value matrix exceeding the target area chip grayscale image are regarded as zero;

[0071] After equalizing the pixel points of the grayscale image of the chip in the target area, each pixel point is smoothed by the Gaussian kernel formula, which can effectively make the change of the pixel point and the adjacent pixel points in a smooth state, so that the pixel points on the chip surface are in a stable change degree when they spread to the surrounding points.

[0072] Step 3: Smooth pixel value HP based on the pixel i after filtering i , traverse the pixel points in the grayscale image of the target area chip, and smooth the pixel value HP of pixel point i i Diffusion deviation value CCp i calculate;

[0073] The step of calculating the diffusion deviation value of the point smoothing pixel value of the pixel point includes:

[0074] Step W1: Get the point smoothing pixel value HP of pixel i i , and at the same time obtain the point smoothed pixel values ​​of the m adjacent pixels of pixel i, and mark them as HP in turn (i,m) ;

[0075] Step W2: Pass Calculate the point smoothing pixel value HP of pixel i i The point smoothing pixel value HP with adjacent pixels (i,m) Deviation value DP(i,m) ;

[0076] Step W3: Calculate the standard deviation of the deviation between the point smoothed pixel value of pixel point i and the point smoothed pixel values ​​of m adjacent pixels to obtain the diffusion deviation value CCp of the point smoothed pixel value of pixel point i i ;

[0077] Step 4: Diffusion deviation value CCp of point smoothing pixel value based on pixel i i ; Determine whether the pixel points of the grayscale image of the chip in the target area are abnormal;

[0078] Specifically, the diffusion deviation value CCp of the point smoothing pixel value of pixel i is i Compare with the diffusion bias threshold CCpy;

[0079] If the diffusion deviation value CCp of the point smoothing pixel value of pixel i i If it is greater than the diffusion deviation threshold CCpy, a pixel abnormal signal is generated;

[0080] Based on the abnormal pixel signal, it is indicated that the chip surface at the pixel in the grayscale image of the chip in the target area is abnormal;

[0081] If the diffusion deviation value CCp of the point smoothing pixel value of pixel i i Less than or equal to the diffusion deviation threshold CCpy, generating a normal pixel signal;

[0082] The diffusion deviation value is obtained by performing difference calculation on the point smoothed pixel value on the pixel point after mapping equalization and filtering smoothing and the point smoothed pixel value of the adjacent pixel point, and then judging whether the diffusion deviation value is within the diffusion deviation threshold, it is judged that the transition change between the pixel point and the adjacent pixel point is smooth and stable, so as to judge whether there is an abnormality at the position of the pixel point on the chip surface, and the calculation comparison is more accurate by performing enhanced equalization and filtering smoothing on the chip grayscale image acquired by X-ray, thereby reducing the misjudgment rate of detection and improving the detection effect and accuracy. Embodiment 2

[0083] Reference Figure 2 As shown, a chip surface defect detection system includes:

[0084] Data acquisition module: used to obtain the chip grayscale image of the target area by X-ray detection, and then mark the pixel points in the chip grayscale image of the target area as i, where i is 1, 2, 3, etc.; at the same time, grayscale value determination is performed based on the total value L of the standard grayscale level to obtain the point grayscale value H of the pixel point in the chip grayscale image of the target area i ; For each pixel i, the gray value H iPerform grayscale mapping enhancement to obtain the point mapping grayscale value GX i ;

[0085] Data processing module: used to map the gray value GX to each pixel i i Perform filtering and obtain the point smoothing pixel value HP i ;

[0086] Deviation calculation module: based on the point smoothing pixel value HP of pixel i after filtering i , traverse the pixel points in the grayscale image of the target area chip, and smooth the pixel value HP of pixel point i i Diffusion deviation value CCp i calculate;

[0087] Abnormal comparison module: Diffusion deviation value CCp of point smoothing pixel value based on pixel i i ; The diffusion deviation value CCp of the point smoothing pixel value of pixel i i Compare with the diffusion deviation threshold CCpy; determine whether the pixel points of the grayscale image of the chip in the target area are abnormal, and obtain the pixel abnormality signal;

[0088] The diffusion deviation value is obtained by performing difference calculation on the point smoothed pixel value on the pixel point after mapping equalization and filtering smoothing and the point smoothed pixel value of the adjacent pixel point, and then judging whether the diffusion deviation value is within the diffusion deviation threshold, it is judged that the transition change between the pixel point and the adjacent pixel point is smooth and stable, so as to judge whether there is an abnormality at the position of the pixel point on the chip surface, and the calculation comparison is more accurate by performing enhanced equalization and filtering smoothing on the chip grayscale image acquired by X-ray, thereby reducing the misjudgment rate of detection and improving the detection effect and accuracy.

[0089] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A chip surface defect detection method, characterized in that: The following steps are involved: Step 1: Obtain a grayscale image of the chip in the target area using X-ray detection, and then mark the pixel points in the grayscale image of the chip in the target area as i, where i is 1, 2, 3, etc.; at the same time, perform grayscale value determination based on the total value L of the standard grayscale level, and obtain the point grayscale value H of the pixel point in the grayscale image of the chip in the target area i ; For each pixel i, the gray value H i Perform grayscale mapping enhancement to obtain the point mapping grayscale value GX i ; Step 2: Map the gray value GX to each pixel i i Perform filtering and obtain the point smoothing pixel value HP i ; Step 3: Smooth pixel value HP based on the pixel i after filtering i , traverse the pixel points in the grayscale image of the target area chip, and smooth the pixel value HP of pixel point i i Diffusion deviation value CCp i calculate; The steps for calculating the diffusion deviation value of the point smoothing pixel value of the pixel point include: Step W1: Get the point smoothing pixel value HP of pixel i i , and at the same time obtain the point smoothed pixel values ​​of the m adjacent pixels of pixel i, and mark them as HP in turn (i,m) ; Step W2: Pass Calculate the point smoothing pixel value HP of pixel i i The point smoothing pixel value HP with adjacent pixels (i,m) Deviation value DP (i,m) ; Step W3: Calculate the standard deviation of the deviation between the point smoothed pixel value of pixel point i and the point smoothed pixel values ​​of m adjacent pixels to obtain the diffusion deviation value CCp of the point smoothed pixel value of pixel point i i ; Step 4: Diffusion deviation value CCp of point smoothing pixel value based on pixel i i ; The diffusion deviation value CCp of the point smoothing pixel value of pixel i i Compare with the diffusion deviation threshold CCpy; determine whether the pixel points of the grayscale image of the chip in the target area are abnormal, and obtain the pixel abnormality signal.

2. A chip surface defect detection method according to claim 1, characterized in that: In the step 1, the gray value H of each pixel i is i Perform grayscale mapping enhancement to obtain the point mapping grayscale value GX i , including the following steps: Based on each pixel point i and its corresponding point grayscale value, the point grayscale value frequency statistics are performed; the point grayscale value frequency set {DH n , r n }; Among them, DH n is the gray value of the point in the chip image of the target area, n is 1, 2, 3...; and n≤i; DH n As n is arranged from small to large; r n DH is the gray value of the point n Number of occurrences; Based on the gray value of the point and the gray value frequency of the gray value of the point, a cumulative pixel value set {DH n , L.J. n }; Among them, LJ n Refers to the grayscale frequency r from the minimum grayscale value to the current grayscale value n The sum of; specifically, ; At this time, based on the pixel i to pixel gray level mapping, the total gray level value DL is defined, where: ; Based on the definition of the total gray level value DL, Calculate the gray value DH of each pixel n Mapping updated point mapping gray value GX (DH n ); Then we get the gray value H of each pixel i i Grayscale mapping enhancement obtains the point mapping grayscale value GX i .

3. A chip surface defect detection method according to claim 1, characterized in that: In the step 2, the gray value GX is mapped to the point i The steps of filtering process include: B01: Constructing the Gaussian kernel formula of kXk , where the standard deviation δ is a preset value; B02: Pass Calculate the normalization factor of the Gaussian kernel formula to obtain the normalization factor value ω of the pixel value filtering; B03: Based on the Gaussian kernel formula and the normalization factor value, traverse the point mapping gray value GX of each pixel in the gray image of the chip in the target area i , filter the point mapping gray value of each pixel i to obtain the point smoothing pixel value HP i .

4. A chip surface defect detection method according to claim 3, characterized in that: In B01, k is greater than or equal to 3.

5. A chip surface defect detection method according to claim 3, characterized in that: In B03, the point mapping grayscale value of each pixel i is filtered to obtain the point smoothing pixel value HP i ; The method is: Take pixel i as the center point, get the pixel matrix kXk, get the point mapping gray value matrix GXgs kXk with pixel i as the midpoint i ; pass Map the gray value GX to the pixel i i Perform filtering calculation to obtain the point smoothing pixel value HP i , where G(x, y) is the Gaussian kernel; GXgs i The gray value matrix of the point kXk centered at pixel i.

6. A chip surface defect detection method according to claim 3, characterized in that: When performing traversal filtering on the point mapping grayscale values ​​of the pixels, if the pixels in the target area chip grayscale image are located at the edge of the target area chip grayscale image, the point mapping grayscale value matrix will exceed the boundary of the pixels in the target area chip grayscale image; at this time, the portion exceeding the boundary is filled with zeros, that is, the pixels in the point mapping grayscale value matrix that exceed the target area chip grayscale image are regarded as zero.

7. A chip surface defect detection method according to claim 1, characterized in that: The diffusion deviation value CCp of the point smoothing pixel value of pixel i i Compare with the diffusion bias threshold CCpy; If the diffusion deviation value CCp of the point smoothing pixel value of pixel i i If it is greater than the diffusion deviation threshold CCpy, a pixel abnormal signal is generated; If the diffusion deviation value CCp of the point smoothing pixel value of pixel i i Less than or equal to the diffusion deviation threshold CCpy, a normal pixel signal is generated.

8. A chip surface defect detection method according to claim 1, characterized in that: Based on the abnormal signal of the pixel point, it is prompted that the chip surface at the pixel point in the grayscale image of the chip in the target area is abnormal.

9. A chip surface defect detection system, characterized in that: The chip surface defect detection system is used to perform the chip surface defect detection method as described in any one of claims 1 to 8, comprising: Data acquisition module: used to obtain the chip grayscale image of the target area by X-ray detection, and then mark the pixel points in the chip grayscale image of the target area as i, where i is 1, 2, 3, etc.; at the same time, grayscale value determination is performed based on the total value L of the standard grayscale level to obtain the point grayscale value H of the pixel point in the chip grayscale image of the target area i ; For each pixel i, the gray value H i Perform grayscale mapping enhancement to obtain the point mapping grayscale value GX i ; Data processing module: used to map the gray value GX to each pixel i i Perform filtering and obtain the point smoothing pixel value HP i ; Deviation calculation module: based on the point smoothing pixel value HP of pixel i after filtering i , traverse the pixel points in the grayscale image of the target area chip, and smooth the pixel value HP of pixel point i i Diffusion deviation value CCp i calculate; Abnormal comparison module: Diffusion deviation value CCp of point smoothing pixel value based on pixel i i ; The diffusion deviation value CCp of the point smoothing pixel value of pixel i i Compare with the diffusion deviation threshold CCpy; determine whether the pixel points of the grayscale image of the chip in the target area are abnormal, and obtain the pixel abnormality signal.