成都大学学报（自然科学版）2025，Vol.44Issue(2)：176-183,8.DOI:10.3969/j.issn.1004-5422.2025.02.009
面向形状特征的模板匹配算法
Template Matching Algorithm for Shape Features
摘要
Abstract
A template matching algorithm based on shape features is proposed to address the issues of low efficiency and poor accuracy in image matching in industrial production in complex environments.Firstly,in the contour template establishment stage,in order to ensure the matching accuracy and efficiency,a self-adaptive search angle step method is introduced.At the same time,a block-level strong binary detec-tion mode based on weak classification is adopted.By utilizing the grayscale characteristics of the image and the block level integration map idea,non-target areas with similar number of contour points to the tar-get image are quickly excluded;secondly,in the contour template matching stage,a pyramid image multi-level matching method is adopted to search for the target pose from coarse matching to fine matching in or-der to reduce the matching time;finally,experimental tests are conducted on four types of workpieces un-der different conditions.The experimental results show that the proposed template matching algorithm can effectively cope with interference such as changes in light intensity and rotation transformation.The matc-hing accuracy of high-quality workpieces reaches over 93.24％,and that of poor-quality workpieces rea-ches over 87.81％.At the same time,the template matching time is less than 594 ms.This algorithm can meet the requirements of high matching efficiency and robustness for image matching.关键词
形状特征/模板匹配/图像金字塔/图像处理/梯度信息Key words
shape features/template matching/image pyramid/image processing/gradient information分类
信息技术与安全科学引用本文复制引用
吕秋荣,孙付春,吴昊荣,李晓晓,徐海..面向形状特征的模板匹配算法[J].成都大学学报（自然科学版）,2025,44(2):176-183,8.基金项目
四川省区域创新合作项目(2024YFHZ0147) （2024YFHZ0147）
