摘要
针对存在非刚性变化的目标识别,提出基于骨架形状特征的目标识别算法。该算法首先提取目标轮廓的骨架,以骨架接合点作为特征点,通过构造与其相连的骨架枝上的骨架特征不变量描述接合点;然后构建一种相似性度量函数对骨架接合点进行匹配,实现目标形状识别。实验结果表明,算法对目标的相似性变换具有较好的稳定性,在目标存在非刚性变化的情况下能够保持较高的识别准确率。
For target recognition with nonrigid changes,a target recognition algorithm based on skeleton shape features was proposed.Firstly the algorithm extracted the skeleton of the target contour,took the skeleton joints as the feature points,and described the joints by constructing the skeleton feature invariants on the skeleton branches connected to it.Secondly,the algorithm constructed a similarity measurement function to match the skeleton joints to realize the recognition of the target shape.The experimental results show that the algorithm in this paper has good stability for the rotation,translation,and zooming of the target,and can maintain a high recognition accuracy when a non-rigid change takes place in the target.
作者
郑伟
于洋
刘砚菊
ZHENG Wei;YU Yang;LIU Yanju(Shenyang Ligong University,Shenyang 110159,China)
出处
《沈阳理工大学学报》
CAS
2022年第1期14-19,共6页
Journal of Shenyang Ligong University
基金
国家重点研发计划基金资助项目(2017YFC0821001)
辽宁省教育厅科学研究项目(LG202009)。
关键词
骨架
接合点
特征不变量
目标识别
skeleton
junction points
feature invariants
target recognition