基于车种查询的轮胎特征信息化研究  

作　　者：梁帅[1,2] 张娜 赵凯[3] 马家骏[3] 

机构地区：[1]铁道警察学院郑州市刑事科学技术重点实验室 [2]郑州大学公共安全研究院 [3]铁道警察学院

出　　处：《道路交通科学技术》2021年第4期22-26,共5页Road Traffic Science & Technology

基　　金：河南省科技攻关计划项目(项目编号:202102310291)阶段性成果;公安部计划研究项目(202QJSXIC29)阶段性成果。

摘　　要：为还原道路交通事故案件真相,确定肇事车辆信息,设计一种车辆轮胎特征自动比对查询系统。整合图像平均值法灰度化、图像直方图均衡化、Otsu阈值(二值化)图像处理和Surf特征点匹配算法,通过特征匹配比对实验进行方案验证,寻找出理想的轮胎花纹图像处理方案,结合方案自建图像特征检索匹配系统,实现轮胎花纹特征自动检索匹配。结果表明轮胎花纹图像灰度化+直方图均衡化+0tsu阈值(二值化)处理的方案具有用时短和匹配识别率高的优点,是一个高效且准确的理想处理方案。In order to restore the truth of road traffic accidents and determine the accident vehicle information,an automatic comparison and query system for vehicle tire characteristics is designed.The authors integrate image average value method grayscale,image histogram equalization,Otsu threshold(binarization)image processing and Surf feature point matching algorithm,verify the scheme through feature matching and comparison experiments and find out an ideal tire pattern image processing scheme.Combined with the self-built image feature retrieval and matching system of the scheme,automatic retrieval and matching of tire pattern features is realized.The results show that the scheme of tire pattern image grayscale+histogram equalization-fOtsu threshold processing(binarization)has the advantages of short time and high matching recognition rate and is an efficient and accurate ideal processing scheme.

关 键 词：轮胎特征 图像处理 特征匹配:信息化 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]