基于图像分割的芯片定位在检测系统中的应用

Application of Chip Location Based on Image Segmentation in Detection System

针对辐射环境下电路板检测对芯片进行定位的情况,提出基于图像分割的芯片定位算法。该算法首先通过人机交互选择待检测芯片并设定初始种子点;其次根据初始种子点进行SLIC超像素合并获得芯片的初始轮廓;再次利用One Cut分割出芯片的环形边缘;最后通过矩形拟合矫正芯片轮廓并获得芯片坐标信息。解决了大尺度的超像素在进行区域合并时,孔洞和断连区域以及跨边缘的超像素导致的轮廓提取不理想的问题。实验结果表明仅需很少的人机交互,可以有效提高图像分割质量,且定位精度在5pixel以内,响应时间在0.8s以内,满足电路板抗辐射检测领域的需求。

Aiming at the situation of chip location in the radiation environment,the paper proposed a chip positioning algorithm based on image segmentation.Firstly,the algorithm selects the test chip and sets the initial seed point by human-computer interaction;secondly,it combines the SLIC super-pixels according to the initial seed points to obtain the initial contour of the chip;thirdly,the loop edge of the chip is again segmented by One Cut;finally,it corrects the chip contour by rectangular fitting and obtains the coordinate information of the chip.The algorithm solved the unsatisfactory contour extraction problem caused by the super-pixel merging of the super-pixel merging region and the super-pixel across the edge.The experimental results show that the algorithm can improve the image segmentation quality effectively with a few human-computer interaction,and the positioning accuracy is less than 5 pixels,and the response time is less than 0.8 s,which meets the needs of the field of radiation resistance detection of circuit boards.