基于SURF算法的汽车底盘异物检测  

Foreign Body Identification of Chassis Based on SURF

作　　者：王粉蝶 普杰信[1] 

机构地区：[1]河南科技大学电子信息工程学院,河南洛阳471003

出　　处：《电视技术》2014年第15期277-280,共4页Video Engineering

基　　金：国家国际科技合作计划(2011DFR10480);河南省基础与前沿计划(102300410113);河南省科技攻关计划(092102210293);国家863计划(2011AA041001)

摘　　要：针对汽车底盘数字图像数据量大、计算耗时长这一难题,构建了汽车底盘数字图像智能异物识别总体结构,提出并实现了基于SURF算法的汽车底盘异物检测。首先利用SURF算法提取特征点并进行匹配,根据匹配得到的单应性矩阵对待测图进行校正,得到标准图与校正图的差分图,然后对差分图进行增强对比度、阈值处理和边缘跟踪,最后标记处异物的位置和大小。实验结果表明,所提出的算法能有效检测出汽车底盘所携带的异物,并能达到系统实时性的要求。An overall structure of car chassis intelligent foreign body identification system is constructed to deal with large data amount and long time consuming in automotive chassis digital images. Foreign objection detection of car chassis based on SURF algorithm is proposed and implemented. Firstly, feature detection and pre-matching of images are done by using the SURF algorithm. Image correction can be finished according to the homographic matrix. Then difference figure by minusing standard image to the correction image is processed by enhancing the contrast, threshold processing and edge tracking, which finally resulting in marking the size and locating of the foreign objection. The experimental results show that the proposed algorithm can effectively detect the foreign body carried by automobile chassis, and can meet the requirement of real-time system.

关 键 词：异物检测 图像配准 SURF算法 阈值处理 边缘追踪 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]