期刊文献+

基于Matlab的鞋底损伤特征分类算法

Matlab-based Algorithm to Classify Shoe-soles’Damaged Characteristics
下载PDF
导出
摘要 目的对鞋底经常出现的三角形、线/曲线、圆形、不规则形四种随机损伤特征进行准确分类,以便后续对鞋底损伤特征的量化研究。方法提出了每个特征的形状由三角形度、线性度、圆形度构成的三维向量决定,对三角形度、线性度、圆形度设置适当的阈值,在分类识别时三维向量的元素和设定的阈值比较得出特征的具体形状。结果利用Matlab软件编制了三角形度、线性度、圆形度的计算程序和阈值比较程序,运行结果表明在提前设定的阈值下通过比较每个形状的三种形度可以实现对形状特征的分类识别,与人工分类识别结果的对比体现了计算机识别的准确率,也体现了计算机识别的定量化优势。结论基于Matlab编写的分类识别程序能够达到正确区分损伤特征和量化损伤特征的目的。 Objective There are four common characteristic damages appearing randomly in the forms of triangle,line/curve,circle and irregularity with the soles of shoes,implying high value and research significance for identification of shoes’soleprints.This paper tries to accurately classify them so as to establish one quantitative approach to the soles’damaged characteristics.Methods Each of the four characteristic damaged forms(triangle,line/curve,circle and irregularity)was defined with the proposed conceptual three-dimensional vectors of triangularity,linearity and circularity.With setting the appropriate threshold of each constituent element of the vectors,Matlab software was selected to calculate the achieving extent of triangularity,linearity and circularity from each damaged form so that the four different damaged characteristics were gauged.Results Through Matlab software to compile the calculation and comparison programs for determination of triangularity,linearity and circularity,the results from the programs performing showed that the four damaged characteristics were quantitatively classified into one certain specific form of them.Such a program-specified comparison excels what human attains in both the accuracy and quantification,especially in the recognition of triangular characteristics.However,limitations were also present from the program-specified comparison with the line/curve identification,showing lower score for the curve.Conclusion Matlab-based classification program can correctly both distinguish and quantify the damaged characteristics.
作者 张庆时 汤澄清 和焕胤 ZHANG Qingshi;TANG Chengqing;HE Huanyin(Criminal Investigation Police University of China,Shenyang 110035,China;Criminal Science and Technology Research Institute of Yangpu Branch of Shanghai Public Security Bureau,Shanghai 200090,China)
出处 《刑事技术》 2020年第4期371-375,I0005,I0006,共7页 Forensic Science and Technology
基金 中国刑事警察学院研究生创新能力提升项目(2019YCZD03)。
关键词 痕迹学 形状特征 阈值 MATLAB 分类识别 traceology characteristic form threshold Matlab classification and recognition
  • 相关文献

参考文献6

二级参考文献33

共引文献109

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部