Feature Identification of Scale-like Scraping Mark on Cars Based on Image Processing
Nowadays,the research on scratching mark identification of cars focuses on paint existence de-tection and collision site confirmation in the practice of traffic accident treatment,and the image features of scale-like scratching mark caused in accident lack of in-depth discussion.Firstly,the layered structure and coating characteristics of car paint were analyzed,and the formation mechanism of the scale-like scratching mark was illustrated.Then,the images from typical accident were collected and processed,and an improved algorithm based on EDCircle was applied to delve the detailed features of the images.Based on the selected cases,when the range of the included angles was set in the range[10,30),the recognition accuracy rate could beyond 90％with a loss rate of less than 30％.Finally,with the verifica-tion by crash cases,theory of scale-like scratching mark mentioned was of high capacity to reconstruct the accident scene,which was a powerful measure of identification to strongly support the identification of ac-cidents.
NETL
NSTL
万方数据
