摘要
在介绍了传统模板匹配算法原理的基础上,分析了该算法在识别和定位目标方面的不足之处,并对其进行了改进,提出将Hu不变矩的模板匹配方法应用在显微视觉目标识别中.实验结果表明:采用Hu不变矩的模板匹配算法解决了传统方法存在的对目标的旋转无法识别的问题,在目标识别上具有旋转和平移不变性,提高了目标识别的准确性.
Based on the introduction of the principle of template matching algorithm,the weakness about it within object recognition and location was analyzed,then the algorithm was improved.Template matching based on Hu invariant moment was applied on target recognition in micro-vision.The result shows that it can overcome the problem that the previous method is sensitive to rotation and improve the accuracy of target recognition with rotation and translation invariance.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2015年第S1期148-151,共4页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家高科技研究发展计划资助项目(2008AA8041302)
国家自然科学基金资助项目(60275013)
关键词
显微视觉伺服
模板匹配
匹配方法
HU不变矩
目标识别
micro-vision servo
template matching
matching method
Hu invariant moment
target recognition
