基于车底阴影前方运动车辆检测.pdfVIP

第 26 卷 第 1 期 电子测量与仪器学报 Vol. 26 No. 1 · 54 · JOURNAL OF ELECTRONIC MEASUREMENT AND INSTRUMENT 2012 年 1 月 DOI: 10.3724/SP.J.1187.2012.00054 基于车底阴影的前方运动车辆检测* 1,2 1 1 齐美彬 潘 燕 张银霞 (1. 合肥工业大学计算机与信息学院, 合肥 230009; 2. 安全关键工业测控技术教育部工程研究中心, 合肥 230009) 摘 要: 实时准确检测前方运动车辆的位置信息是车辆安全驾驶的前提, 提出了一种在不依赖车道线检测情况下, 基于车 底阴影的前方运动车辆检测。阴影算法采用两次自适应阈值分割图像从而提取车底与路面的交线, 生成目标假设区域, 对建筑 物或树的投影、光照强弱等干扰能初步有效的排除, 提高检测效率; 接着, 利用熵值归一化的对称性测度来验证, 排除虚假车 辆。实验结果显示, 该算法的正确识别率达到 97.2%, 平均处理速度为 22.5 帧/秒, 对白天中多种环境能满足实时准确地要求。 关键词: 前方车辆; 检测; 阴影; 对称性 中图分类号: TP391 文献标识码: A 国家标准学科分类代码: 580.2010 Preceding moving vehicle detection based on shadow of chassis Qi Meibin1,2 Pan Yan1 Zhang Yinxia1 (1. School of Computer and information, Hefei University of Technology, Hefei 230009, China; 2. Engineering Research Center of Safety Critical Industrial Measurement and Control Technology, Hefei 230009, China) Abstract: Real time detection of the location information of the preceding vehicle is the premiss for security driving. A detection method for preceding vehicles based on the shadow of chassis without lane detection is presented. In order to generate hypothetical areas, the shadow algorithm adopted two adaptive thresholds to extract the intersecting lines of chassis and the road, and could eliminate the false areas such as the shadow of buildings or trees and improve the detection efficiency. Then, a target vehicle was affirmed by entropy-normalized-symmetry for rejecting false vehicles. The experimental results show that the propo