Chip defect detection method based on three-dimensional point cloud grid reconstruction

Abstract

The application discloses a chip defect detection method based on three-dimensional point cloud grid reconstruction, comprising the following steps: 1, collecting three-dimensional point cloud of a reference chip; 2, denoising; 3, ordered sampling; 4, grid processing: constructing a square network, and then dividing each square into two triangular facets; 5, grid optimization: adjusting and optimizing the height coordinates of each triangular facet by minimizing the average distance of all three-dimensional point clouds in each square to the corresponding triangular facet, and then obtaining an optimized three-dimensional grid model; 6, collecting three-dimensional point cloud of a target part of a target chip; and 7, chip defect detection. The application is based on a three-dimensional grid model, can effectively detect the surface defects (such as missing, warping, deviation and the like) of the chip, can quickly locate the pin defects (such as excessive tin, insufficient tin and the like), has high automation degree and detection precision, and is particularly suitable for chip manufacturing and quality control processes with high precision and high quality requirements.

Description

Technical Field

[0001] This invention relates to the field of chip defect detection in the semiconductor industry, and in particular to a chip defect detection method based on three-dimensional point cloud mesh reconstruction. Background Technology

[0002] With the rapid development of the semiconductor industry, the complexity and precision requirements of chip manufacturing are increasing. Defect detection is a crucial step in ensuring product quality and reliability during chip production. Traditional defect detection methods are mostly based on two-dimensional image processing technology, using optical microscopes or X-rays to inspect the chip surface. While these methods can effectively detect certain surface defects, their reliance on two-dimensional images often makes it difficult to identify complex three-dimensional defects, especially for detecting minute defects, defects with uneven depth, or structural defects such as pins.

[0003] The limitations of two-dimensional inspection technology have spurred the application of three-dimensional vision technology in chip defect detection. Three-dimensional inspection methods acquire complete three-dimensional data of the chip's surface and internal structure, enabling a more intuitive and comprehensive analysis of chip defects, and offering higher detection precision and accuracy. Traditional three-dimensional inspection technologies typically rely on methods such as laser scanning and CT scanning. While these technologies can generate high-quality three-dimensional data, they are complex to operate, costly, and have limited real-time detection capabilities in production environments. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to address the shortcomings of the prior art by providing a chip defect detection method based on three-dimensional point cloud mesh reconstruction. This chip defect detection method based on three-dimensional point cloud mesh reconstruction can realize real-time automatic detection of chip body height defects and pin soldering defects.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0006] A chip defect detection method based on 3D point cloud mesh reconstruction includes the following steps.

[0007] Step 1: Collect 3D point cloud of the reference chip: Collect the corresponding 3D point cloud for each part of the reference chip that needs to be inspected for 3D defects; one part corresponds to one 3D point cloud set.

[0008] Step 2, Denoising: Denoising is performed on each 3D point cloud acquired in Step 1.

[0009] Step 3, Ordered Sampling: For each denoised 3D point cloud, ordered sampling is performed to form a corresponding ordered point cloud.

[0010] Step 4: Mesh processing: Using all ordered sampling points in each ordered point cloud set as the corner points of the grid, construct a grid network, and then divide each grid into two triangular facets to form a preliminary three-dimensional mesh model; wherein, the three vertices of each triangular facet are the three ordered sampling point clouds in the corresponding ordered point cloud set.

[0011] Step 5, Mesh Optimization: By minimizing the average distance from all denoised 3D point clouds in each square to the corresponding triangular facet, the height coordinates of each triangular facet are adjusted and optimized to obtain an optimized 3D mesh model.

[0012] Step 6: Acquire 3D point cloud of target chip and target part: Acquire and denoise the 3D point cloud of the target part in the target chip to be tested.

[0013] Step 7, Chip Defect Detection: By retrieving the optimized 3D mesh model corresponding to the target part, high-level defect detection can be performed on the target chip.

[0014] In step 3, the denoised 3D point cloud set P out The method for performing ordered sampling includes the following steps:

[0015] Step 3-1: Obtain the extreme values ​​of point cloud coordinates: Traverse the 3D point cloud set P out The extreme values ​​of point cloud coordinates in the X and X directions are calculated as follows: maximum value of point cloud coordinates in the X direction Xmax, minimum value of point cloud coordinates in the X direction Xmin, maximum value of point cloud coordinates in the Y direction Ymax, and minimum value of point cloud coordinates in the Y direction Ymin.

[0016] Step 3-2: Set the sampling step size: For the 3D point cloud set P out Set the X-axis sampling step size t respectively x and Y-axis sampling step size t y .

[0017] Step 3-3: Divide the sampling area: Divide the 3D point cloud P out Divide the region into m regions along the X direction and n regions along the Y direction, where:

[0018]

[0019] Step 3-4: Finding ordered sampling points: In the middle of each sampling region divided in step 3-3, select one point as the found ordered sampling point.

[0020] In step 3-3, the 3D point cloud set P outThe i-th intersection region along the X direction and the j-th intersection region along the Y direction is denoted as the sampling region P(i,j). At this time, the sampling region P(i,j) contains N(i,j) denoised 3D point clouds, namely: p1, p2, p3, ..., p N(i,j) In steps 3-4, the ordered sampling points found within the sampling region P(i,j) are denoted as s(i,j), and the X, Y, and Z coordinates of s(i,j) are denoted as s(i,j).x, s(i,j).y, and s(i,j).z, respectively. The formulas for calculating s(i,j).x and s(i,j).y are as follows:

[0021] s(i,k).x=Xmin+(i-1)*t x

[0022] s(i,j).y=Ymin+(j-1)*t y

[0023] s(i,j).z is determined based on the number of N(i,j), specifically as follows:

[0024] D. When N(i,j)=0, and i=1 or j=1, the specific calculation formula is as follows:

[0025] s(i,j).z=Zstd

[0026] In the formula, Zstd is the height of the reference plane, which is obtained by performing plane fitting on the 3D point cloud of the part acquired in step 1;

[0027] E. When N(i,j)=0, i≠1 and j≠1, the specific calculation formula is as follows:

[0028]

[0029] In the formula, s(p,q).z is the Z-coordinate of the ordered sampling point corresponding to the p-th intersection region along the X direction and the q-th intersection region along the Y direction; where p takes the values ​​i-1, i, and i+1; and q takes the values ​​j-1, j, and j+1.

[0030] N(p,q) represents the number of denoised 3D point clouds contained in the p-th intersection region along the X direction and the q-th intersection region along the Y direction.

[0031] N is the total number of N(p,q) > 0;

[0032] The sum of the Z-axis coordinates of all ordered sampling points in the region N(p,q)>0;

[0033] F. When N(i,j)≥1, the specific calculation formula is as follows:

[0034]

[0035] In the formula, p1.z, p2.z, p3.z, ..., p N(i,j) .z represents the 3D point cloud p1, p2, p3, ..., p N(i,j) The Z-coordinate.

[0036] Step 4, the method for constructing the preliminary 3D mesh model, includes the following steps:

[0037] Step 4-1: Find the corner points of the grid: Take the ordered sampling point s(i,j) as the starting corner point of a grid A, and search for the neighboring ordered sampling points to obtain the other three corner points of grid A, namely: s(i+1,j), s(i,j+1), s(i+1,j+1).

[0038] Step 4-2: Construct triangular facets: Using s(i,j) in square A as the starting point, construct two triangular facets.

[0039] Step 4-3: Constructing the preliminary 3D mesh model: In Step 3-4, a total of m*n ordered sampling points are obtained. Steps 4-1 to 4-2 are repeated, and each ordered sampling point is traversed, resulting in (m-1)*(n-1) squares, thus obtaining a preliminary 3D mesh model M containing 2*(m-1)*(n-1) triangular facets. in .

[0040] In step 4-2, the two triangular faces of square A are V1(i,j) and V2(i,j), which can be constructed in two ways:

[0041] Construction method 1: V1(i,j)={s(i,j),s(i+1,j),s(i+1,j+1)},V2(i,j)={s(i,j),s(i,j+1),s(i+1,j+1)}。

[0042] Construction method 2: V1(i,j)={s(i,j),s(i+1,j),s(i,j+1)},V2(i,j)={s(i+1,j),s(i,j+1),s(i+1,j+1)}。

[0043] Step 5, the mesh optimization method, includes the following steps:

[0044] Step 5-1: Set the minimum distance threshold Thres for the facets.

[0045] Step 5-1: Extract 3D point clouds: The denoised 3D point clouds corresponding to the two triangular faces V1(i,j) and V2(i,j) in grid A are denoised as 3D point cloud sets P1(i,j) and P2(i,j), respectively; where P1(i,j) = {p11,p12,p13…p1n1}, P2(i,j) = {p21,p22,p23…p2n2}; where n1 and n2 are the number of 3D point clouds corresponding to the triangular faces V1(i,j) and V2(i,j), respectively.

[0046] Step 5-2, Calculate d(i,j): d(i,j) is the average distance from the 3D point cloud in grid A to the corresponding triangular facet. The specific calculation formula is as follows:

[0047]

[0048] in:

[0049]

[0050] In the formula, d1(i,j) is the average distance from all three-dimensional point clouds in the three-dimensional point cloud set P1(i,j) to the triangular patch V1(i,j).

[0051] These are the distances from the first, second, third, and n1th 3D point clouds in the 3D point cloud set P1(i,j) to the triangular facet V1(i,j), respectively.

[0052] d2(i,j) is the average distance from all 3D point clouds in the 3D point cloud set P2(i,j) to the triangular patch V2(i,j).

[0053] These are the distances from the first, second, third, and n2th 3D point clouds in the 3D point cloud set P2(i,j) to the triangular facet V2(i,j), respectively.

[0054] Step 5-3: Calculate the distance D: D is the average value of d(i,j) across all squares. The specific calculation formula is as follows:

[0055]

[0056] Step 5-4, Mesh Optimization: Traverse each cell in the initial 3D mesh model and calculate its corresponding d(i,j), then compare it with the minimum distance threshold Thres for the facets. Specifically:

[0057] A. When d(i,j)<Thres, the ordinates of the three corner points of the two triangular faces in the corresponding square do not need to be adjusted.

[0058] B. When d(i,j)≥Thres, adjust the z-coordinates of the three corner points of the two triangular faces in the corresponding grid, and recalculate d(i,j) and D using the least squares method; when D is minimized, the optimized 3D mesh model M is obtained. out .

[0059] In step 5-4, when adjusting the z-coordinates of the three corner points of the triangular facet, the z-coordinates of the ordered sampling points s(1,j), s(m,j), s(i,1) and s(i,n) located in the bounding box remain unchanged.

[0060] In step 7, the target chip height type defects include chip body height defects and pin soldering defects; among them, chip body height defects include chip body height offset defects and chip body warping defects.

[0061] Step 7, the method for detecting target chip height-type defects, includes the following steps:

[0062] Step 7-A1: Create four detection areas: In the optimized 3D mesh model, using the reference edge of the reference chip or target chip's border, offset the four corners of the reference chip or target chip inward by L and H respectively to form four detection areas R1, R2, R3 and R4 with the same set size; where L is the set length offset value and H is the set width offset value.

[0063] Step 7-A2, Calculate the average height of the reference chip: Calculate the average height of the reference chip in the four detection regions R1, R2, R3, and R4 of the optimized 3D mesh model, which are respectively: H model1 H model2 H model3 H model4 .

[0064] Step 7-A3: Calculate the height threshold: Based on the average height of the reference chip obtained in Step 7-A2, set the height threshold; wherein, the height threshold includes the minimum height threshold H. min and maximum height threshold H max The specific expression is:

[0065] Hmin = scale * (H model1 +H model2 +H model3 +H model4 ) / 4

[0066] Hmax = threshold + (H model1 1+H model2 +H model3 +H model4 ) / 4

[0067] In the formula, scale is the minimum height threshold ratio coefficient, and scale∈(0,1).

[0068] thres is the base setting for the maximum height threshold.

[0069] Step 7-A4: Set the length direction threshold H len、 and width direction threshold H wid .

[0070] Step 7-A5: Calculate the average height of the target chip: Calculate the average height values ​​of the target chip in the four detection regions R1, R2, R3 and R4 of the optimized 3D mesh model, which are H1, H2, H3 and H4 respectively.

[0071] Step 7-A6: Target chip height type defect detection. The specific detection expression is as follows:

[0072]

[0073] When H1, H2, H3, and H4 obtained in step 7-A5 simultaneously satisfy the above five expressions, the target chip body has no height defects; otherwise, the target chip body has height defects, specifically:

[0074] A. If the values ​​of H1, H2, H3, and H4 are all less than H... min This indicates that the target chip itself has a component missing defect.

[0075] B. If any one of H1, H2, H3, and H4 is less than H... min The remaining height range is within the interval (H) min H max If the value is 0, it indicates that the target chip itself has a high offset defect;

[0076] C. If any one of H1, H2, H3, and H4 is greater than H... max ,

[0077] Or Abs(H1-H2)>H wid

[0078] Or Abs(H3-H4)>H wid

[0079] Or Abs(H1-H3)>H len

[0080] Or Abs(H2-H4)>H len

[0081] This indicates that the target chip itself has a warping defect.

[0082] Step 7, the method for detecting solder joint defects, includes the following steps:

[0083] Step 7-B1: Create pin detection areas: In the optimized 3D mesh model, create pin detection areas R of the same set size for the soldering part of each pin of the reference chip or target chip.

[0084] Step 7-B2: Calculate the extreme values ​​of the average height of the reference chip: The extreme values ​​of the average height of the reference chip include the minimum extreme value of the average height, Hmod. min and the maximum average height extreme value Hmod max Among them, Hmod min Hmod represents the minimum average height in the detection area R of all pins of the reference chip. max The average maximum height is the maximum value of the detection area R of all pins of the reference chip.

[0085] Step 7-B3: Set the height threshold H0.

[0086] Step 7-B4: Calculate the extreme value of the target chip height: Calculate the minimum height H of the target chip in all pin detection areas R of the optimized 3D mesh model. min and maximum height H max .

[0087] Step 7-B5, Pin soldering defect detection: When H min <Hmod min When -H0, the pin has insufficient solder; when H max >Hmod max When +H0, the pin has excessive solder.

[0088] The present invention has the following beneficial effects:

[0089] 1. This invention presents a chip defect detection method based on 3D point cloud reconstruction. This method collects and processes point cloud data from the chip surface to generate a precise 3D mesh model, and then performs chip height defect analysis based on this model. Compared to traditional 2D methods, 3D point cloud reconstruction technology provides richer spatial information, thus more accurately identifying defects on the chip surface and pins, such as missing parts, misalignment, warping, excessive solder, and insufficient solder. It has significant advantages, especially in detecting excessive or insufficient solder on small defects, complex shapes, and pins, enabling higher precision defect localization and analysis.

[0090] 2. This invention also has a high level of automation, enabling real-time defect monitoring and quality control during chip manufacturing, reducing human inspection errors and improving production efficiency.

[0091] 3. This invention is based on 3D point cloud mesh reconstruction technology, which can effectively eliminate noise at the 3D imaging end and perform more accurate 3D mesh reconstruction, achieving precise 3D mesh reconstruction of the chip. In other words, 3D point cloud mesh reconstruction technology has brought a significant technological breakthrough to the field of chip defect detection, meeting the demands for high precision, high efficiency, and automation. Attached Figure Description

[0092] Figure 1 The flowchart of a chip defect detection method based on three-dimensional point cloud mesh reconstruction according to the present invention is shown.

[0093] Figure 2 This diagram shows a grid formed by ordered sampling points during the initial three-dimensional mesh model construction process of this invention.

[0094] Figure 3 The figures show two examples of constructing a preliminary three-dimensional mesh model according to the present invention; where (a) and (b) correspond to two preliminary three-dimensional mesh models, respectively.

[0095] Figure 4 A schematic diagram of chip body detection is shown; where (a) and (b) correspond to two different shapes of chip bodies, respectively.

[0096] Figure 5 A schematic diagram of chip pin detection is shown; where (a) and (b) correspond to two different pin shapes, respectively. Detailed Implementation

[0097] The present invention will now be described in further detail with reference to the accompanying drawings and specific preferred embodiments.

[0098] In the description of this invention, it should be understood that the terms "left side," "right side," "upper part," "lower part," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. "First," "second," etc., do not indicate the importance of the components, and therefore should not be construed as a limitation of this invention. The specific dimensions used in this embodiment are only for illustrating the technical solution and do not limit the scope of protection of this invention.

[0099] like Figure 1 As shown, a chip defect detection method based on three-dimensional point cloud mesh reconstruction includes the following steps.

[0100] Step 1: Acquire 3D point cloud of the reference chip: For each part in the reference chip that needs to be inspected for 3D defects, its corresponding 3D point cloud is acquired by scanning with multiple sensors; one part corresponds to one 3D point cloud set.

[0101] Step 2, Denoising: Denoising is performed on each 3D point cloud acquired in Step 1, specifically including the following steps.

[0102] Step 2-1: Perform plane fitting on the input point cloud P from Step 1 to obtain the reference plane height Z = Zstd.

[0103] Step 2-2: Set the minimum height threshold t1 and the maximum height threshold t2.

[0104] Steps 2-3: Extract the point cloud within the interval [Zstd-t1, Zstd+t2] by axial clipping to form a new point cloud P. out .

[0105] Step 3, Ordered Sampling: For each denoised 3D point cloud, ordered sampling is performed to form a corresponding ordered point cloud.

[0106] The above describes the denoised 3D point cloud set P out The method for performing ordered sampling includes the following steps.

[0107] Step 3-1: Obtain the extreme values ​​of point cloud coordinates: Traverse the 3D point cloud set P out The extreme values ​​of point cloud coordinates in the X and Y directions are calculated as follows: maximum value of point cloud coordinates in the X direction Xmax, minimum value of point cloud coordinates in the X direction Xmin, maximum value of point cloud coordinates in the Y direction Ymax, and minimum value of point cloud coordinates in the Y direction Ymin.

[0108] Step 3-2: Set the sampling step size: For the 3D point cloud set P out Set the X-axis sampling step size t respectively x and Y-axis sampling step size t y .

[0109] Step 3-3: Divide the sampling area: Divide the 3D point cloud P out Divide the region into m regions along the X direction and n regions along the Y direction, where:

[0110]

[0111] The above will be the three-dimensional point cloud P out The i-th region along the X direction and the j-th region along the Y direction are denoted as sampling region P(i,j). At this time, sampling region P(i,j) contains N(i,j) denoised 3D point clouds, namely: p1, p2, p3, ..., p... N(i,j) .

[0112] Step 3-4: Finding ordered sampling points: In the middle of each sampling region divided in step 3-3, select one point as the found ordered sampling point.

[0113] The ordered sampling points found within the sampling region P(i,j) are denoted as s(i,j), and the X, Y, and Z coordinates of s(i,j) are denoted as s(i,j).x, s(i,j).y, and s(i,j).z, respectively. The formulas for calculating s(i,j).x and s(i,j).y are as follows:

[0114] s(i,j).x=Xmin+(i-1)*t x

[0115] s(i,j).y=Ymin+(j-1)*t y

[0116] s(i,j).z is determined based on the number of N(i,j), specifically as follows:

[0117] A. When N(i,j)=0, and i=1 or j=1, the specific calculation formula is as follows:

[0118] s(i,j).z=Zstd

[0119] In the formula, Zstd is the height of the reference plane, which is obtained by performing plane fitting on the 3D point cloud of the part acquired in step 1;

[0120] B. When N(i,j)=0, i≠1 and j≠1, the specific calculation formula is as follows:

[0121]

[0122] In the formula, s(p,q).z is the Z-coordinate of the ordered sampling point corresponding to the p-th intersection region along the X direction and the q-th intersection region along the Y direction; where p takes the values ​​i-1, i, and i+1; and q takes the values ​​j-1, j, and j+1.

[0123] N(p,q) represents the number of denoised 3D point clouds contained in the p-th intersection region along the X direction and the q-th intersection region along the Y direction.

[0124] N is the total number of N(p,q) > 0;

[0125] The sum of the Z-axis coordinates of all ordered sampling points in the region N(p,q)>0;

[0126] C. When N(i,j)≥1, the specific calculation formula is as follows:

[0127]

[0128] In the formula, p1.z, p2.z, p3.z, ..., p N(i,j) .z represents the 3D point cloud p1, p2, p3, ..., p N(i,j) The Z-coordinate.

[0129] Step 4: Mesh processing: Using all ordered sampling points in each ordered point cloud set as the corner points of the grid, construct a grid network, and then divide each grid into two triangular facets to form a preliminary three-dimensional mesh model; wherein, the three vertices of each triangular facet are the three ordered sampling point clouds in the corresponding ordered point cloud set.

[0130] Step 5, Mesh Optimization: By minimizing the average distance from all denoised 3D point clouds in each square to the corresponding triangular facet, the height coordinates of each triangular facet are adjusted and optimized to obtain an optimized 3D mesh model.

[0131] The method for constructing the above-mentioned preliminary three-dimensional mesh model preferably includes the following steps.

[0132] Step 4-1, Locate the corner points of the grid: such as Figure 2 and Figure 3 As shown, the ordered sampling point s(i,j) is taken as the starting corner point of a square A, and the neighboring ordered sampling points are searched to obtain the other three corner points of square A, namely: s(i+1,j), s(i,j+1), s(i+1,j+1).

[0133] Step 4-2: Construct triangular facets: Using s(i,j) in square A as the starting point, construct two triangular facets. For example... Figure 3 As shown, the two triangular faces of square A are V1(i,j) and V2(i,j), which can be constructed in two ways:

[0134] Construction method 1: V1(i,j)={s(i,j),s(i+1,j),s(i+1,j+1)},V2(i,j)={s(i,j),s(i,j+1),s(i+1,j+1)}。

[0135] Construction method 2: V1(i,j)={s(i,j),s(i+1,j),s(i,j+1)},V2(i,j)={s(i+1,j),s(i,j+1),s(i+1,j+1)}。

[0136] Step 4-3: Constructing the preliminary 3D mesh model: In Step 3-4, a total of m*n ordered sampling points are obtained. Steps 4-1 to 4-2 are repeated, and each ordered sampling point is traversed, resulting in (m-1)*(n-1) squares, thus obtaining a preliminary 3D mesh model M containing 2*(m-1)*(n-1) triangular facets. in .

[0137] The above-mentioned grid optimization method preferably includes the following steps.

[0138] Step 5-1: Set the minimum distance threshold Thres for the facets.

[0139] Step 5-1: Extract 3D point clouds: The denoised 3D point clouds corresponding to the two triangular faces V1(i,j) and V2(i,j) in grid A are denoised as 3D point cloud sets P1(i,j) and P2(i,j), respectively; where P1(i,j) = {p11,p12,p13…p1n1}, P2(i,j) = {p21,p22,p23…p2n2}; where n1 and n2 are the number of 3D point clouds corresponding to the triangular faces V1(i,j) and V2(i,j), respectively.

[0140] Step 5-2, Calculate d(i,j): d(i,j) is the average distance from the 3D point cloud in grid A to the corresponding triangular facet. The specific calculation formula is as follows:

[0141]

[0142] in:

[0143]

[0144] In the formula, d1(i,j) is the average distance from all three-dimensional point clouds in the three-dimensional point cloud set P1(i,j) to the triangular patch V1(i,j).

[0145] These are the distances from the first, second, third, and n1th 3D point clouds in the 3D point cloud set P1(i,j) to the triangular facet V1(i,j), respectively.

[0146] d2(i,j) is the average distance from all 3D point clouds in the 3D point cloud set P2(i,j) to the triangular patch V2(i,j).

[0147] These are the distances from the first, second, third, and n2th 3D point clouds in the 3D point cloud set P2(i,j) to the triangular facet V2(i,j), respectively.

[0148] Step 5-3: Calculate the distance D: D is the average value of d(i,j) across all squares. The specific calculation formula is as follows:

[0149]

[0150] Step 5-4, Mesh Optimization: Traverse each cell in the initial 3D mesh model and calculate its corresponding d(i,j), then compare it with the minimum distance threshold Thres for the facets. Specifically:

[0151] A. When d(i,j)<Thres, the ordinates of the three corner points of the two triangular faces in the corresponding square do not need to be adjusted.

[0152] B. When d(i,j)≥Thres, adjust the z-coordinates of the three corner points of the two triangular faces in the corresponding grid, and recalculate d(i,j) and D using the least squares method; when D is minimized, the optimized 3D mesh model M is obtained. out .

[0153] When adjusting the z-coordinates of the three corner points of the triangular facet, the z-coordinates of each edge-ordered sampling point s(1,j), s(m,j), s(i,1), and s(i,n) located in the bounding box remain unchanged.

[0154] This invention effectively suppresses cusp noise and solves the noise fluctuation caused by high brightness in components by adjusting the z-value of the ordinate of each ordered sampling point in the preliminary three-dimensional mesh model and adjusting the triangular facets of the model using the least squares method.

[0155] Step 6: Acquire 3D point cloud of target chip and target part: Acquire and denoise the 3D point cloud of the target part in the target chip to be tested.

[0156] Step 7, Chip Defect Detection: By retrieving the optimized 3D mesh model corresponding to the target part, high-level defect detection can be performed on the target chip.

[0157] The aforementioned target chip height type defects include chip body height defects and pin soldering defects; among them, chip body height defects include chip body height offset defects and chip body warping defects.

[0158] The detection method for the aforementioned target chip height-type defects includes the following steps.

[0159] Step 7-A1: Create four detection regions: In the optimized 3D mesh model, using the reference edge of the reference chip or target chip's border, offset the four corners of the reference chip or target chip inward by L and H respectively, forming regions as shown in the figure. Figure 4 The four detection areas R1, R2, R3, and R4 shown are of the same set size, as detailed below. Figure 4 As shown; where L is the set length offset value; H is the set width offset value.

[0160] Step 7-A2, Calculate the average height of the reference chip: Calculate the average height of the reference chip in the four detection regions R1, R2, R3, and R4 of the optimized 3D mesh model, which are respectively: H model1 H model2 H model3 H model4 .

[0161] Step 7-A3: Calculate the height threshold: Based on the average height of the reference chip obtained in Step 7-A2, set the height threshold; wherein, the height threshold includes the minimum height threshold H. min and maximum height threshold H max The specific expression is:

[0162] Hmin = scale * (H model1 +H model2 +H model3 +H model4 ) / 4

[0163] Hmax = threshold + (H model1 1+H model2 +H model3 +H model4 ) / 4

[0164] In the formula, scale is the minimum height threshold ratio coefficient, and scale∈(0,1).

[0165] thres is the basic setting value of the maximum height threshold. In this embodiment, the height unit is preferably set to micrometers, so thres > 50 micrometers.

[0166] Step 7-A4: Set the length direction threshold H len、 and width direction threshold H wid .

[0167] Step 7-A5: Calculate the average height of the target chip: Calculate the average height values ​​of the target chip in the four detection regions R1, R2, R3 and R4 of the optimized 3D mesh model, which are H1, H2, H3 and H4 respectively.

[0168] Step 7-A6: Target chip height type defect detection. The specific detection expression is as follows:

[0169]

[0170] When H1, H2, H3, and H4 obtained in step 7-A5 simultaneously satisfy the above five expressions, the target chip body has no height defects; otherwise, the target chip body has height defects, specifically:

[0171] A. If the values ​​of H1, H2, H3, and H4 are all less than H... min This indicates that the target chip itself has a component missing defect.

[0172] B. If any one of H1, H2, H3, and H4 is less than H... min The remaining height range is within the interval (H) min H maxIf the value is 0, it indicates that the target chip itself has a high offset defect;

[0173] C. If any one of H1, H2, H3, and H4 is greater than H... max ,

[0174] Or Abs(H1-H2)>H wid

[0175] Or Abs(H3-H4)>H wid

[0176] Or Abs(H1-H3)>H len

[0177] Or Abs(H2-H4)>H len

[0178] This indicates that the target chip itself has a warping defect.

[0179] The above-mentioned method for detecting solder joint defects preferably includes the following steps.

[0180] Step 7-B1: Create Pin Detection Regions: In the optimized 3D mesh model, create pin detection regions R of the same set size for the solder joints of each pin of the reference chip or target chip, specifically as follows: Figure 5 As shown.

[0181] Step 7-B2: Calculate the extreme values ​​of the average height of the reference chip: The extreme values ​​of the average height of the reference chip include the minimum extreme value of the average height, Hmod. min and the maximum average height extreme value Hmod max ; where Hmod min Hmod represents the minimum average height in the detection area R of all pins of the reference chip. max The average maximum height is the maximum value of the detection area R of all pins of the reference chip.

[0182] Step 7-B3: Set the height threshold H0.

[0183] Step 7-B4: Calculate the extreme value of the target chip height: Calculate the minimum height H of the target chip in all pin detection areas R of the optimized 3D mesh model. min and maximum height H max .

[0184] Step 7-B5, Pin soldering defect detection: When H min <Hmod min When -H0, the pin has insufficient solder; when H max >Hmod max When +H0, the pin has excessive solder.

[0185] The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the specific details of the above embodiments. Within the scope of the technical concept of the present invention, various equivalent transformations can be made to the technical solutions of the present invention, and these equivalent transformations all fall within the protection scope of the present invention.

Claims

1. A chip defect detection method based on three-dimensional point cloud mesh reconstruction, characterized in that: Includes the following steps: Step 1: Acquire 3D point cloud of the reference chip: Acquire the corresponding 3D point cloud for each part of the reference chip that needs to be inspected for 3D defects; one part corresponds to one 3D point cloud set. Step 2, Denoising: Denoising is performed on each 3D point cloud acquired in Step 1; Step 3, Ordered Sampling: For each denoised 3D point cloud, ordered sampling is performed to form a corresponding ordered point cloud; Step 4: Mesh processing: Using all ordered sampling points in each ordered point cloud set as the corner points of the grid, construct a grid network, and then divide each grid into two triangular facets to form a preliminary three-dimensional mesh model; wherein, the three vertices of each triangular facet are the three ordered sampling point clouds in the corresponding ordered point cloud set; Step 5, Mesh Optimization: By minimizing the average distance from all denoised 3D point clouds in each square to the corresponding triangular facet, the height coordinates of each triangular facet are adjusted and optimized to obtain an optimized 3D mesh model. Step 6: Acquire 3D point cloud of target chip and target part: Acquire and denoise the 3D point cloud of the target part in the target chip to be inspected; Step 7, Chip Defect Detection: By retrieving the optimized 3D mesh model corresponding to the target part, high-level defect detection can be performed on the target chip; The target chip height type defects include chip body height defects and pin soldering defects; among them, chip body height defects include chip body height offset defects, component missing defects, and chip body warping defects; The method for detecting height-type defects in the target chip includes the following steps: Step 7-A1: Create four detection areas: In the optimized 3D mesh model, using the reference edge of the reference chip or target chip's border, offset the four corners of the reference chip or target chip inward by L and H respectively to form four detection areas R1, R2, R3 and R4 of the same set size; where L is the set length offset value and H is the set width offset value. Step 7-A2, Calculate the average height of the reference chip: Calculate the average height of the reference chip in the four detection regions R1, R2, R3, and R4 of the optimized 3D mesh model, which are respectively: H model1 H model2 H model3 H model4 ; Step 7-A3: Calculate the height threshold: Based on the average height of the reference chip obtained in Step 7-A2, set the height threshold; wherein, the height threshold includes the minimum height threshold H. min and maximum height threshold H max The specific expression is: Hmin= scale *( H model1 + H model2 + H model3 + H model4 ) / 4; Hmax=thres+( H model1 1+ H model2 + H model3 + H model4 ) / 4; In the formula, This is the minimum height threshold scaling factor, and ; thres is the base setting for the maximum height threshold; Step 7-A4: Set the length direction threshold H len、 and width direction threshold H wid ; Step 7-A5: Calculate the average height of the target chip: Calculate the average height of the target chip in the four detection regions R1, R2, R3 and R4 of the optimized 3D mesh model, which are H1, H2, H3 and H4 respectively; Step 7-A6: Target chip height type defect detection. The specific detection expression is as follows: ； When H1, H2, H3, and H4 obtained in step 7-A5 simultaneously satisfy the above five expressions, the target chip body has no height defects; otherwise, the target chip body has height defects, specifically: A. If All four values ​​are less than This indicates that the target chip itself has a component missing defect; B. If In, any value less than The remaining height range is within the interval. This indicates that the target chip itself has a high offset defect; C. If In, any value greater than , or ; or ; or ; or ; This indicates that the target chip itself has a warping defect.

2. The chip defect detection method based on three-dimensional point cloud mesh reconstruction according to claim 1, characterized in that: In step 3, the denoised 3D point cloud set P is processed. out The method for performing ordered sampling includes the following steps: Step 3-1: Obtain the extreme values ​​of point cloud coordinates: Traverse the 3D point cloud set P out Calculations yielded that and The extreme values ​​of the point cloud coordinates in the directions are as follows: Maximum point cloud coordinates , Minimum point cloud coordinates , Maximum point cloud coordinates , Minimum point cloud coordinates ; Step 3-2: Set the sampling step size: For the 3D point cloud set P out Set the X-axis sampling step size respectively and Y-axis sampling step size ; Step 3-3: Divide the sampling area: Divide the 3D point cloud P out along Divide the direction into m regions, along The direction is divided into n regions, where: ； ； Step 3-4: Finding ordered sampling points: In the middle of each sampling region divided in step 3-3, select one point as the found ordered sampling point.

3. The chip defect detection method based on three-dimensional point cloud mesh reconstruction according to claim 2, characterized in that: In step 3-3, the 3D point cloud set P out along The i-th direction, along The j-th intersection region is denoted as the sampling region P(i,j); at this time, the sampling region P(i,j) contains N(i,j) denoised 3D point clouds, which are as follows: In steps 3-4, the ordered sampling points found within the sampling region P(i,j) are denoted as... ,and The X, Y, and Z coordinates are denoted as follows: , and ;but and The calculation formula is: ； ； The value is determined based on the number of N(i,j), specifically as follows: A. When N(i,j)=0, and i=1 or j=1, the specific calculation formula is as follows: ； In the formula, The reference plane height is obtained by performing plane fitting on the 3D point cloud of the part acquired in step 1; B. When N(i,j)=0, i≠1 and j≠1, the specific calculation formula is as follows: ； In the formula, For along The p-th direction, along The Z-coordinate of the ordered sampling point corresponding to the q-th intersection region in the direction; where p takes the value of , and ; q takes the value , and ; For along The p-th direction, along The number of denoised 3D point clouds contained in the qth intersection region of the direction; for The total quantity; For all The sum of the Z-coordinates of the ordered sampling points in the region; C. When N(i,j)≥1, the specific calculation formula is as follows: ； In the formula, 3D point cloud The Z-coordinate.

4. The chip defect detection method based on three-dimensional point cloud mesh reconstruction according to claim 3, characterized in that: Step 4, the method for constructing the preliminary 3D mesh model, includes the following steps: Step 4-1, Find the corner points of the grid: Organize the ordered sampling points Starting from a corner point of square A, we search for neighboring ordered sampling points to obtain the other three corner points of square A, which are: , ; Step 4-2, Constructing Triangular Facets: Using square A as an example... As the starting point, construct two triangular faces; Step 4-3: Constructing the preliminary 3D mesh model: In Step 3-4, a total of [number] mesh models were obtained. For each ordered sampling point, repeat steps 4-1 to 4-2, iterating through each ordered sampling point to obtain a total of [number missing]. Each square contains... A preliminary 3D mesh model M of a triangular facet in .

5. The chip defect detection method based on three-dimensional point cloud mesh reconstruction according to claim 4, characterized in that: In step 4-2, the two triangular faces of square A are V1(i,j) and V2(i,j), which can be constructed in two ways: Construction method 1: V1(i,j)={s(i,j),s(i+1,j),s(i+1,j+1)}, V2(i,j)={s(i,j),s(i,j+1),s(i+1,j+1)}; Construction method 2: V1(i,j)={s(i,j),s(i+1,j),s(i,j+1)}, V2(i,j)={s(i+1,j),s(i,j+1),s(i+1,j+1)}.

6. The chip defect detection method based on three-dimensional point cloud mesh reconstruction according to claim 5, characterized in that: Step 5, the mesh optimization method, includes the following steps: Step 5-1: Set the minimum distance threshold Thres for the facets; Step 5-1: Extract 3D point clouds: The denoised 3D point clouds corresponding to the two triangular faces V1(i,j) and V2(i,j) in grid A are denoised as 3D point cloud sets P1(i,j) and P2(i,j), respectively; where P1(i,j)={p11,p12,p13…p1n1}, P2(i,j)={p21,p22,p23…p2n2}; where n1 and n2 are the number of 3D point clouds corresponding to the triangular faces V1(i,j) and V2(i,j), respectively. Step 5-2, Calculation : The average distance from the 3D point cloud in grid A to the corresponding triangular facet is calculated using the following formula: ； in: ； ； In the formula, Let V1 be the average distance from all 3D point clouds in the 3D point cloud set P1(i,j) to the triangular patch V1(i,j); These are the distances from the first, second, third, and n1th 3D point clouds in the 3D point cloud set P1(i,j) to the triangular facet V1(i,j), respectively. Let V2 be the average distance from all 3D point clouds in the 3D point cloud set P2(i,j) to the triangular patch V2(i,j); These are the distances from the first, second, third, and n2th 3D point clouds in the 3D point cloud set P2(i,j) to the triangular facet V2(i,j), respectively. Step 5-3: Calculate distance D: D is the distance among all squares. The average value is calculated using the following formula: ； Step 5-4, Mesh Optimization: Traverse each cell in the initial 3D mesh model and calculate its corresponding... And compare it with the minimum distance threshold Thres for the facets, specifically: A. When < In this case, the ordinates of the three corner points of the two triangular faces in the corresponding square do not need to be adjusted; B, when ≥ If the result is not found, then the z-coordinates of the three corner points of the two triangular faces in the corresponding square are adjusted, and the least squares method is used to recalculate the result. And D; when D is minimized, the optimized 3D mesh model M is obtained. out .

7. The chip defect detection method based on three-dimensional point cloud mesh reconstruction according to claim 6, characterized in that: In step 5-4, when adjusting the z-coordinates of the three corner points of the triangular facet, the ordered sampling points located on each edge within the bounding box are... , , and The z-value of the ordinate remains unchanged.

8. The chip defect detection method based on three-dimensional point cloud mesh reconstruction according to claim 1, characterized in that: Step 7, the method for detecting solder joint defects, includes the following steps: Step 7-B1: Create pin detection areas: In the optimized 3D mesh model, create pin detection areas R of the same set size for the soldering part of each pin of the reference chip or target chip. Step 7-B2: Calculate the extreme values ​​of the average height of the reference chip: The extreme values ​​of the average height of the reference chip include the minimum extreme value of the average height, Hmod. min and the maximum average height extreme value Hmod max Among them, Hmod min Hmod represents the minimum average height in the detection area R of all pins of the reference chip. max The average maximum height in the detection area R of all pins of the reference chip; Step 7-B3: Set the height threshold H0; Step 7-B4: Calculate the extreme value of the target chip height: Calculate the minimum height H of the target chip in all pin detection areas R of the optimized 3D mesh model. min and maximum height H max ; Step 7-B5, Pin soldering defect detection: When H min <Hmod min - When H0, the pin has insufficient solder; when H max >Hmod max When + H0, the pin has excessive solder.

Images