基于机器视觉的芯片引脚缺陷检测系统设计与实现

针对双球红外接收头芯片人工缺陷检测难度大、误判率高等问题,设计了一个基于机器视觉的引脚缺陷检测系统,对双球红外接收头芯片的引脚进行缺陷检测,达到分辨出合格品和瑕疵品的目的;首先,通过工业相机实时采集芯片图像,并对图像进行滤波、灰度化等预处理;然后利用VisionPro视觉软件的PMAlign工件进行图像特征匹配,计算引脚个数以判断引脚是否缺失,利用AnglePonitPonit工具计算引脚间距以判断引脚是否弯曲;最后将检测到的芯片位置信息和识别结果通过socket通讯协议发送给工业机器人;工业机器人根据识别结果,将合格品和瑕疵品分别抓取至不同区域,实现对芯片的分类管理;实验结果表明,该缺陷检测系统误判率为0.4%,满足工业生产的要求。

Design and implementation of chip pin defect detection system based on machine vision

Aiming at the difficulty and low accuracy of manual defect detection of the dual-ball infrared receiver chip, a pin defect detection system based on machine vision is designed to detect the pins of the dual-ball infrared receiver chip to distinguish normal Chips and defective chips. First, the chip image is collected in real time through an industrial camera, and preprocessing such as filtering and grayscale processing is performed; then the PMAlign tool of VisionPro vision software is used for image feature matching, the number of pins is calculated to determine whether the pins are missing, and the AnglePonitPonit tool is used to calculate Pin spacing to determine whether the pins are bent; Finally, the detected chip position information and recognition results are sent to the industrial robot through the socket communication protocol. According to the recognition result, the industrial robot grabs the normal chips and defective chips to different areas to realize the classification management of the chips. Experimental results show that the detection accuracy of the defect detection system reaches ±0.01mm, and the misjudgment rate is less than 0.5%, which meets the requirements of industrial production.