基于机器视觉的芯片编号识别算法

针对芯片检测过程中芯片编号形状小、人眼无法识别的现状,开发了一套芯片编号识别系统。系统以机器视觉技术为基础,配备高分辨率Basler相机、高倍率镜头以及高性能点光源,能实现高分辨率图像采集;利用Hough直线检测对图像进行水平旋转,利用图像轮廓查找算法对兴趣区域进行提取分割,利用帧差法对兴趣区域与标准库进行比较,找到最匹配的图像,其标签为识别结果。实验结果表明,运用系统的图像处理算法可以代替人眼通过高倍放大镜识别,改进了芯片检测工艺,大大提高了芯片检测的效率。

Recognition algorithm of chip numbers based on machine vision

In the process of chip detection,a set of numbers recognition system has been developed for the current situation that the numbers on the chip are small and the human eye cannot recognize.Based on machine vision technology,the system was equipped with high-resolution Basler camera,high-magnification lens and high-performance point light source to capture high-resolution images.Hough line detection was used to rotate the image horizontally,and image contour search algorithm was used to extract the region of interest.Using temporal difference method to compare the region of interest with the standard library,finding the best matching image whose label was the recognition result.Test result shows that image processing algorithm can replace human eye recognition by high magnification magnifier,which greatly improved the chip detection process and the efficiency of chip detection.