Search

CN-121998969-A - Gold thread segment detection method and system in chip AOI detection, electronic equipment and computer readable medium

CN121998969ACN 121998969 ACN121998969 ACN 121998969ACN-121998969-A

Abstract

The invention discloses a gold wire segment detection method and system in chip AOI detection, electronic equipment and a computer readable medium, wherein the method comprises the following steps of obtaining a chip image, performing welding spot positioning on the chip image to obtain a welding spot coordinate of a gold wire And two solder joint coordinates Calculate the length L and inclination angle of the gold wire based on the coordinates of the first and second solder joints Dividing gold wire into multiple segments, each segment corresponding to a detection area, wherein the detection area is composed of retracted end point connection lines, and the retracted end points are formed according to coordinates of a welding point and two welding points, preset retraction proportion and length And angle of The method comprises the steps of calculating, carrying out threshold segmentation on each gold wire section by using a gray threshold corresponding to the gold wire section, extracting the gold wire section, judging whether the gold wire sections are communicated or not and judging whether the whole gold wire is communicated or not, so as to determine whether the gold wire has defects or not. The invention reduces the complexity and cost of gold thread detection algorithm.

Inventors

  • HAN YAHUI

Assignees

  • 上海铭沣科技股份有限公司

Dates

Publication Date
20260508
Application Date
20260306

Claims (10)

  1. 1. A gold thread segment detection method in chip AOI detection is characterized by comprising the following steps: Acquiring a chip image; Performing welding spot positioning on the chip image to obtain a welding spot coordinate of the gold wire And two solder joint coordinates ; Calculating the length L and the inclination angle of the gold wire according to the coordinates of the first welding point and the second welding point ; Dividing gold wires into a plurality of sections, wherein each section corresponds to a detection area, the detection area is formed by connecting lines of retracted end points, and the retracted end points are formed according to the coordinates of the first welding point and the second welding point, the preset retraction proportion and the length And angle of Calculating to obtain; For each gold wire segment, carrying out threshold segmentation by using a gray threshold corresponding to the gold wire segment, and extracting a gold wire region of the gold wire segment; Judging whether the gold wire areas of each section are communicated or not and whether the whole gold wire is communicated or not so as to determine whether the gold wire has defects or not.
  2. 2. The method for detecting gold wire segments in AOI detection of a chip according to claim 1, wherein the chip image is subjected to solder joint positioning to obtain a solder joint coordinate of gold wire And two solder joint coordinates The method comprises the following steps: threshold segmentation is carried out on the chip image, and a region with gray values within a preset range is extracted to be used as a potential welding spot region; performing open operation on the potential welding spot area to remove noise points; for a welding spot area, calculating the center of an inscribed circle as a welding spot coordinate For the two welding spot areas, calculating the center of gravity as the coordinates of the two welding spots 。
  3. 3. The method for detecting gold wire segments in chip AOI detection according to claim 1, wherein the length L and the inclination angle of the gold wire The calculation formula of (2) is as follows: ; 。
  4. 4. the method for detecting gold thread segments in chip AOI test of claim 1, wherein the retracted end points are calculated according to the following formula: for the retraction of a welding spot direction, the retraction proportion is set as A coordinate of a welding spot after retraction The method comprises the following steps: ; for the retraction of the two welding spot directions, the retraction proportion is set as A coordinate of a welding spot after retraction The method comprises the following steps: 。
  5. 5. The method for detecting gold thread segments in chip AOI detection according to claim 1, wherein the step of judging whether each gold thread region is connected is to judge whether the extracted gold thread region contains two indentation endpoints corresponding to the gold thread region for each gold thread region, if so, the gold thread region is connected, otherwise, the gold thread region is broken or deviated.
  6. 6. The method for detecting gold thread segments in chip AOI detection according to claim 1, wherein the judging whether the whole gold thread is connected specifically comprises: Combining the gold thread regions to obtain a communication region of the whole gold thread; Judging the coordinates of a welding spot before retraction And two solder joint coordinates And if the gold wires are all positioned in the communication area, the gold wires are normal, otherwise, the gold wires are broken or offset.
  7. 7. The method for detecting gold thread segments in chip AOI detection according to claim 1, wherein the gray threshold corresponding to each gold thread segment is independently set according to the actual gray distribution of the gold thread segment in the image.
  8. 8. A gold thread segmentation detecting system in chip AOI detection is characterized by comprising the following steps: The image acquisition module is used for acquiring a chip image; The welding spot positioning module is used for performing welding spot positioning on the chip image to obtain a welding spot coordinate of the gold wire And two solder joint coordinates ; The parameter calculation module is used for calculating the length L and the inclination angle of the gold wire according to the coordinates of the first welding point and the second welding point ; The subsection calculation module is used for dividing the gold wire into a plurality of sections, each section corresponds to a detection area, the detection area is composed of retracted end point connecting wires, and the retracted end points are formed according to the coordinates of the first welding point and the second welding point, the preset retraction proportion and the length And angle of Calculating to obtain; the segmentation extraction module is used for carrying out threshold segmentation by using a gray threshold value corresponding to each section of gold wire, and extracting a gold wire region of the section; the defect judging module is used for judging whether the gold wire areas of each section are communicated or not and whether the whole gold wire is communicated or not so as to determine whether the gold wire has defects or not.
  9. 9. An electronic device comprising a memory, a processor and a computer program, wherein the computer program is stored in the memory, and the processor runs the computer program to execute a method for detecting gold wire segments in chip AOI detection according to any one of claims 1 to 7.
  10. 10. A computer readable medium having non-volatile program code executable by a processor, said program code causing said processor to run a method of chip AOI detection in accordance with any one of claims 1 to 7.

Description

Gold thread segment detection method and system in chip AOI detection, electronic equipment and computer readable medium Technical Field The invention relates to the technical field of semiconductor defect detection, in particular to a gold thread segment detection method and system in chip AOI detection, electronic equipment and a computer readable medium. Background With the continuous evolution of integrated circuit manufacturing processes and the development of packaging technologies toward high density and high performance, wire Bonding (WB) is used as a core process for electrically interconnecting chips, and the quality of the Wire Bonding (WB) directly determines the reliability of the package and the performance of the final electronic device. In WB process, micron-sized metal wires (usually gold wires) connect the chip pads and external pins in a specific space arc, forming a precise path for electrical signal transmission. Therefore, the morphology, position and bonding point integrity of the gold thread are key points for evaluating the quality of WB and preventing potential failure. Automated Optical Inspection (AOI) systems rely on the acquisition and analysis of two-dimensional images of gold wires with high accuracy. However, since the gold wire is a cylindrical structure with a spatial radian, under the irradiation of a single fixed light source, the light intensity reflected to the camera is significantly different due to the different directions of the surface normals of different sections. This physical property causes the same gold wire to exhibit serious gray-scale unevenness phenomenon on a two-dimensional image, such as partial section highlighting, partial section dullness, or even insufficient contrast with the background. Existing detection schemes based on traditional image processing algorithms generally aim to detect a single gold wire as a whole at a time. The core idea is to design or adopt a robust complex hunting algorithm (such as dynamic programming, self-adaptive edge tracking and the like) which can adapt to gray level gradual change so as to directly extract a complete and continuous gold thread outline from the background. However, this "global detection" strategy faces a significant bottleneck in practical applications: 1. The algorithm has high complexity and serious time consumption, and the complicated hunting algorithm needs iterative calculation or global optimization to solve the interruption and interference caused by uneven gray scale, and has huge calculation cost. In the high-density chip packaging image, the processing time is multiplied, and the production efficiency of real-time detection is restricted. 2. The parameters are sensitive, and the stability is difficult to ensure, and the performance of the algorithm is highly dependent on preset initial points, energy function parameters or regularization constraints. In the face of small changes of different products, different illumination conditions or gold thread forms, algorithms may fail (such as hunting diverges and early termination), so that detection results are unstable, and frequent manual intervention and parameter readjustment are required. 3. The engineering implementation and maintenance difficulty is high, the complex algorithm model is not easy to develop and debug, the development and maintenance cost of software is increased, and the operability of a user is reduced. Disclosure of Invention According to a first aspect of the embodiment of the invention, there is provided a gold wire segment detection method in chip AOI detection, comprising the steps of: Acquiring a chip image; Performing welding spot positioning on the chip image to obtain a welding spot coordinate of the gold wire And two solder joint coordinates; Calculating the length L and the inclination angle of the gold wire according to the coordinates of the first welding point and the second welding point; Dividing gold wires into a plurality of sections, wherein each section corresponds to a detection area, the detection area is formed by connecting retracted end points, and the retracted end points are formed according to coordinates of a welding point and a welding point, a preset retraction proportion and a preset lengthAnd angle ofCalculating to obtain; For each gold wire segment, carrying out threshold segmentation by using a gray threshold corresponding to the gold wire segment, and extracting a gold wire region of the gold wire segment; Judging whether the gold wire areas of each section are communicated or not and whether the whole gold wire is communicated or not so as to determine whether the gold wire has defects or not. Further, the chip image is subjected to welding spot positioning to obtain a welding spot coordinate of the gold wireAnd two solder joint coordinatesThe method comprises the following steps: threshold segmentation is carried out on the chip image, and a region with gray values within a preset range