Search

CN-122023899-A - Pogpin life monitoring system based on visual detection

CN122023899ACN 122023899 ACN122023899 ACN 122023899ACN-122023899-A

Abstract

The application relates to the technical field of semiconductor automatic test equipment, and discloses a pogpin life monitoring system based on visual detection, which comprises an image detection module, an image preprocessing module, a processor, a server and an early warning module, wherein the image detection module acquires appearance images before and after pogpin contact by using upper and side double-station acquisition units, the image preprocessing module integrates graying, RC filtering and self-adaptive threshold segmentation circuits, hardware-level denoising and enhancing are carried out on the image, the processor extracts the needle point outline and calculates Euclidean distance of the needle point vertex before and after testing as a deformation quantity, the early warning level is generated based on comparison of the deformation quantity and the threshold, and the server stores data and drives the early warning module to execute hierarchical warning through a 5G network. The application eliminates the detection blind area, reduces the operation load of the processor, and realizes the real-time quantitative monitoring and automatic closed-loop early warning of the pogpin service life.

Inventors

  • Shang Anhang
  • Peng Sanmao
  • YANG PEIFAN

Assignees

  • 南京伟测半导体科技有限公司

Dates

Publication Date
20260512
Application Date
20260128

Claims (10)

  1. 1. A pogpin life monitoring system based on visual detection, comprising: The image detection module is used for acquiring appearance images of the pogopin before and after the pogopin contacts the chip; The input end of the image preprocessing module is connected with the image detection module and is used for denoising and characteristic enhancement processing of the appearance image so as to output a processed image signal; The processor is connected with the output end of the image preprocessing module and is used for receiving the processed image signals, extracting the outline characteristics of the needle tip, calculating the deformation quantity of the needle tip before and after the pogpin test, and comparing the needle tip deformation quantity with a preset threshold value to generate an early warning grade; The server is connected with the processor and used for storing the needle point deformation and the early warning grade; And the early warning module is connected with the server and is used for executing the warning action according to the early warning grade.
  2. 2. The vision detection-based pogpin life monitoring system of claim 1, wherein the image detection module comprises a first image acquisition unit and a second image acquisition unit; The first image acquisition unit is arranged right above the initial station and is used for capturing an appearance image before pogopin is contacted with the chip; the initial station is a position where the needle card moves to the upper part of the chip and does not execute the pressing action; the second image acquisition unit is arranged at the side of the contact station and is used for capturing an appearance image when the pogopin is contacted with the chip; The contact station is the position where the pogopin and the chip pins form a conducting loop after the pin card performs the pressing action.
  3. 3. The vision detection-based pogpin life monitoring system of claim 1, wherein the image preprocessing module is internally integrated with a graying processing unit, an RC filtering unit and an adaptive threshold segmentation unit; The graying processing unit is used for converting the original color image signals corresponding to the appearance images into gray image signals; The RC filter unit is connected to the output end of the graying processing unit and is used for carrying out smoothing processing on the gray image signals; The adaptive threshold segmentation unit is connected to the output end of the RC filter unit and is used for calculating a local threshold and performing binarization segmentation on the smoothed image signal.
  4. 4. A pogpin life monitoring system based on visual inspection according to claim 3, wherein the graying processing unit is configured to obtain gray values by calculating a weighted sum of red channel component, green channel component and blue channel component by a weighted average method; the self-adaptive threshold segmentation unit is configured to obtain the local threshold by calculating the average value of gray values of all pixels in a neighborhood window taking the current pixel point as the center and subtracting a preset trimming constant.
  5. 5. The vision detection-based pogpin life monitoring system of claim 1, wherein the processor is internally provided with a feature recognition circuit and a needle point variable calculation logic circuit; The feature recognition circuit is used for extracting pinpoint outline data and pinpoint vertex coordinate data of the pogopin from the processed image signals; The needle point variable computing logic circuit is used for computing the needle point deformation based on the needle point vertex coordinate data.
  6. 6. The vision detection-based pogpin life monitoring system of claim 5, wherein the feature recognition circuit is configured to construct a needle point contour point set by screening pixels with non-zero gradients, and select an extreme point of the needle point contour point set in a vertical direction as a needle point vertex; the needle tip variable calculation logic is configured to calculate a euclidean distance between a needle tip vertex coordinate in an initial state and a needle tip vertex coordinate in an operating state, and take the euclidean distance as the needle tip variable.
  7. 7. The vision detection-based pogpin life monitoring system of claim 1, wherein the image detection module is connected with the image preprocessing module through an HDMI interface; the image preprocessing module is connected with the processor through an RS485 interface; the processor is connected with the server through an RS485 interface; The server is internally provided with an industrial grade 5G communication module, and the server is connected with the early warning module through a wireless communication link.
  8. 8. The pogpin life monitoring system based on visual detection of claim 1, wherein the early warning module comprises an audible and visual alarm unit and a machine working interface popup window unit; the audible and visual alarm unit is arranged at the top of the detection machine and consists of a red light, a yellow light, a green light and an industrial buzzer; The machine working interface popup window unit is integrated in an operating system of the detection machine.
  9. 9. The vision detection-based pogpin life monitoring system of claim 8, wherein a hardware register built in the processor stores a life early warning threshold and a life limit threshold, the life early warning threshold being smaller than the life limit threshold; when the needle point deformation is smaller than the service life early warning threshold, the early warning module drives the green light to be normally on; when the needle point deformation is larger than or equal to the service life early warning threshold value and smaller than the service life limit threshold value, the early warning module drives the yellow lamp to flash, the industrial buzzer to sound in a gap, and drives the machine working interface popup window unit to pop up prompt information; When the needle point deformation is greater than or equal to the service life limit threshold, the early warning module drives the red light to be always on, the industrial buzzer continuously sounds, and drives the machine working interface popup window unit to pop up warning information.
  10. 10. The vision-based pogpin life monitoring system of claim 1, wherein the system further comprises a feedback terminal; The feedback terminal is an industrial touch screen and is connected with the server through a wireless communication link; The feedback terminal is used for receiving and displaying the needle point deformation and the early warning level, and responding to an operation instruction of a user to derive historical data.

Description

Pogpin life monitoring system based on visual detection Technical Field The invention relates to the technical field of semiconductor automatic test equipment, in particular to a pogpin life monitoring system based on visual detection. Background In a semiconductor wafer test (CP) or package test (FT) link, pogpin (probe) is used as a key medium for connecting a test machine and a chip pin, and its physical state directly affects the accuracy and yield of a test result. In the long-term high-frequency contact test process of pogopin, the pinpoint can be worn, bent or deformed, once the pinpoint is abnormal, the contact impedance is unstable to cause false measurement if the pinpoint is light, and the wafer is scrapped if the pinpoint is heavy. Current pogpin status monitoring techniques rely mainly on manual periodic off-line inspection or simple on-line vision systems. In terms of image acquisition, existing monitoring means generally use a single-view industrial camera or microscope for observation, and focus on static overhead detection in a non-contact state. However, failure of pogon pin often accompanies abnormal bending or lateral sliding in the stress process, and a single overlooking angle is difficult to cover lateral morphology change of pogon pin in the contact pressing process, so that a visual blind area exists in detection of the stress deformation of a needle body, and early tiny deformation characteristics are difficult to comprehensively capture. In terms of image data processing, the existing visual detection system generally adopts a framework of 'camera acquisition-industrial personal computer processing', namely, a front-end camera transmits an original high-resolution image to a rear-end main control computer, and the main control computer completes all preprocessing work such as denoising, filtering, segmentation and the like through a software algorithm. Under the scene that the electromagnetic environment of the test machine is complex and the production beat is performed at high speed, the pure software processing mode occupies a large amount of transmission bandwidth and CPU operation resources, so that the system processing delay is higher, and the timeliness requirement of online real-time monitoring is difficult to meet. In terms of early warning decision mechanism, the existing maintenance strategy mostly adopts preventive replacement with fixed test times or relies on reverse push investigation after the test yield is obviously reduced. The method lacks of real-time quantitative monitoring of physical deformation of pogopin, and cannot provide accurate digital early warning before the tiny deformation of the needle tip occurs but batch test accidents are not caused yet. The existing detection means can only provide qualitative qualification/disqualification judgment, and the defect of grading early warning logic based on deformation degree causes obvious hysteresis in maintenance work, so that quality risk in the production process is increased. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a pogpin service life monitoring system based on visual detection, which solves the technical problems of high operation load and poor real-time performance caused by the dependence on software processing and maintenance lag caused by the lack of physical deformation quantization indexes in single visual angle detection in the prior art. In order to achieve the above purpose, the invention is realized by the following technical scheme: the invention provides a pogpin life monitoring system based on visual detection, which comprises an image detection module, an image preprocessing module, a processor, a server and an early warning module. The image detection module is used for obtaining appearance images of the pogopin before and after the pogopin contacts the chip. The input end of the image preprocessing module is connected with the image detection module and is used for denoising and characteristic enhancement processing of the appearance image so as to output a processed image signal. The processor is connected with the output end of the image preprocessing module, and is used for receiving the processed image signals, extracting the outline characteristics of the needle tip, calculating the deformation quantity of the needle tip before and after the pogpin test, and comparing the needle tip deformation quantity with a preset threshold value to generate an early warning grade. The server is connected with the processor and used for storing the needle point deformation and the early warning grade. The early warning module is connected with the server and used for executing the warning action according to the early warning grade. Further, the image detection module is composed of a first image acquisition unit and a second image acquisition unit. The first image acquisition unit is arranged right above an initial station and is used for capturing