Research And Development On Visual Detection Method And Sorting System For Recyclable Chips

In recent years,the domestic chip demand is increasing,and the "chip shortage" has lasted for a long time.Therefore,it is of great value to renovate and utilize recyclable chips.In the process of chip disassembly,problems such as solder connection on the surface,scratches and missing affect chip reuse.In order to meet the intelligent detection demand of chip recycling enterprises for recyclable chips,an intelligent system is developed to realize automatic detection of surface defects of recycled chips and automatic sorting according to their size.The main contents are as follows:（1）According to the design requirements of chip detection and sorting system,the design of system operation,detection and lighting scheme is completed,which is composed of vibration plate feeding system,glass rotary plate control system and double optical acquisition system,and the selection of hardware in the optical acquisition system,including camera,lens,light source,etc.（2）According to the requirements of system operation,single-thread control scheme and multi-thread control scheme are designed and analyzed in terms of motion control algorithm,and multi-thread control scheme is selected after comprehensive comparison.The algorithm of this scheme has high efficiency and stable operation,and can realize flying beat detection.The theoretical maximum detection efficiency can reach 5pcs/s.（3）As for the image processing algorithm,station 1 uses the method of edge fitting and image enhancement algorithm to complete the location of the chip,and obtains the surface area of the chip by the output parameters of the function in the location algorithm,and completes the sorting of the chip by the difference of the surface area.Blob analysis was used in station 2 to detect tin connection,tin deficiency and scratch defects of chips.This method was used to identify the defective chips by comparing the solder area and solder quantity on the chip surface,and the comprehensive recognition rate was 96.5%.（4）the use of c # computer language to motion control algorithm in the human-computer interaction and visual algorithm embedded system software,completed the design of the software and programming,the characteristics of its parameters can be configured for different specifications chips for unique set of parameters,and applied to the rotating disk sorting machine vision,after the test,obtaining good results.