火力与指挥控制2025，Vol.50Issue(4)：120-127,8.DOI:10.3969/j.issn.1002-0640.2025.04.017
改进区域灰度模型的晶圆针印边缘亚像素检测
Sub-pixel Detection of Wafer Probe Print Edge Based on Improved Regional Grayscale Model
摘要
Abstract
To achieve high-precision detection of the wafer probe print image region's edge,this paper proposes a subpixel detection method based on an improved regional grayscale model.First,anisotropic diffusion filtering is used to replace gaussian filtering in the canny operator,enabling pixel-level edge localization of the probe print region.Then,a 9×3 sub-image is established at the edge points,and an adaptive weighting method is designed to determine the weight values of overlapping pixels.The overlapping pixels are weighted during sub-image merging,and a new sub-image is constructed through iteration.Finally,sub-pixel points are calculated based on regional grayscale features in the sub-image to obtain precise coordinates.Experimental results show that the improved algorithm can achieve high-precision edge detection for wafer probe prints.The localization accuracy is within 0.25 pixels,and in the case of mixed gaussian noise,the localization accuracy is better than that of canny-d,the traditional zernike moments,and regional grayscale model methods.lt also has strong noise resistance,stability,and good processing efficiency.关键词
亚像素/晶圆针印/区域灰度模型/边缘检测/自适应加权Key words
sub-pixel/wafer probe print/regional grayscale model/edge detection/adaptive weighting分类
信息技术与安全科学引用本文复制引用
仲重光,马国军,朱勤华,张博文,王亚军,杨俊杰..改进区域灰度模型的晶圆针印边缘亚像素检测[J].火力与指挥控制,2025,50(4):120-127,8.基金项目
国家自然科学基金(61371114) （61371114）
江苏省研究生科研与实践创新计划基金资助项目(SJCX23_2141) （SJCX23_2141）
