摘要
针对被部分遮挡车辆的漏检率高这一难点,在深入分析可变形部件模型的基础上,提出了基于双车辆可变形部件模型的车辆检测算法。该算法采用对图像分区域匹配和对匹配结果进行融合的方法来减少多车辆检测中被部分遮挡车辆检测的漏检情况。实验结果表明:该算法在部分遮挡车辆的检测中要优于已有算法,它明显地降低了漏检率,满足安全驾驶辅助技术应用中的实时性要求。
In view of the difficult issue of high miss detection rate in detecting partially obscured vehicles, on the basis of in-depth analysis on deformable part model,a vehicle detection algorithm based on double-vehicle deformable part model is proposed. In the algorithm a method of image segmentation matching and matching results fusion is adopted to reduce the miss detection rate of partially obscured vehicles in multi-vehicle detection. Experi-mental results show that the algorithm proposed is superior to the existing algorithms in partially obscured vehicle de-tection ,obviously lowering the miss detection rate, meeting the real-time requirements in the application of safe driving assistance technology.
出处
《汽车工程》
EI
CSCD
北大核心
2017年第6期710-715,721,共7页
Automotive Engineering
基金
国家自然科学基金(61403172
61601203
U1564201)
中国博士后基金(2014M561592
2015T80511)
江苏省六大人才高峰项目(2015-JXQC-012
2014-DZXX-040)
江苏省自然科学基金(BK20140555)资助
关键词
汽车工程
驾驶辅助技术
多车辆检测
可变形部件模型
遮挡车辆检测
automotive engineering
driving assistance technology
multi-vehicle detection
deformable part model
obscured vehicle detection