融合小波变换与形态学的车身刮痕边缘检测  

Edge Detection of Car Body Scratches Based on Wavelet Transform and Morphology

作　　者：杨智勇 刘研奇 刘佳豪 缪陈 YANG Zhi-yong;LIU Yan-qi;LIU Jia-hao;MIAO Chen(School of Mechanical Engineering,Hubei University of Technology,Hubei Wuhan 430068,China;95997 Troops of the Chinese People’s Liberation Army,Beijing 100071,China;State Grid Hu’nan Electric Power Co.,Ltd.,Maintenance Company,Hu’nan Changsha 410004,China)

机构地区：[1]湖北工业大学机械工程学院,湖北武汉430068 [2]中国人民解放军95997部队,北京100071 [3]国网湖南省电力有限公司检修公司,湖南长沙410004

出　　处：《机械设计与制造》2024年第2期54-59,共6页Machinery Design & Manufacture

基　　金：国家自然科学基金资助项目(51907055)。

摘　　要：针对车身表面外部噪声干扰的影响,导致无法准确提取到车身刮痕边缘有效信息,提出一种融合小波变换和数学形态学的边缘检测方法。首先,为减弱车身刮痕边缘提取过程中受光照等外界噪声干扰影响,对刮痕源图像利用小波模极大值边缘检测方法,消除部分噪声干扰,提高边缘检测定位的准确性;然后,为了优化刮痕边缘的连续性,避免在多尺度小波检测下边缘有效信息的丢失,提出一种改进的数学形态学边缘检测算法;最后,将小波模极大值与数学形态学两种方法的检测结果进行小波分解,分别得到图像的高频和低频分量。两图像低频分量根据图像能量大小分配权重进行融合,对应的高频分量通过绝对值取大的方式融合,将融合后的高频和低频分量重构,获取最终的刮痕边缘。通过实验分析,该融合方法能保证在去噪的同时,获得更完善清晰的边缘特征信息,效果优于传统边缘检测算法,能为后续刮痕精确定位修复提供理论与技术支持。In view of the influence of strong local reflection and external noise interference on the surface of the car body,it is impossible to accurately extract the effective information of the car body scratch edge.It proposes an edge detection method that combines wavelet transform and mathematical morphology.First of all,in order to reduce the influence of external noise interference such as light in the process of car body scratch edge extraction,the wavelet modulus maximum edge detection method is used for the scratch source image to eliminate some noise interference and improve the accuracy of edge detection and positioning;Optimize the continuity of the scratched edge to avoid the loss of effective edge information under multi-scale wavelet detection,and propose an improved mathematical morphology edge detection algorithm;finally,the detection of wavelet modulus maximum and mathematical morphology are two methods.The low-frequency components of the two images are fused according to the weight of the image energy distribution to obtain the fused low-frequency component.The corresponding high-frequency component is compared with the absolute value,and the larger of the two is used as the high-frequency component of the fused image.High frequency and low frequency components are reconstructed to obtain the final scratch edge.Through experimental analysis,the fusion method can guarantee to obtain more complete and clear edge feature information while removing noise,the effect is better than traditional edge detection algorithms,and it can provide theoretical and technical support for the subsequent precise location and repair of scratches.

关 键 词：小波变换 数学形态学 边缘检测 汽车刮痕 

分 类 号：TH16[机械工程—机械制造及自动化] TH878