基于多通道背景提取算法的车辆检测
2018-07-28周盛王宵王子豪陶菁怡邵叶秦李泽慧胡彬陈锦花王琼
周盛 王宵 王子豪 陶菁怡 邵叶秦 李泽慧 胡彬 陈锦花 王琼
摘要:车辆检测是智能交通中的一个基本问题。为了有效地检测车辆,本文提出了基于多通道背景提取算法的车辆檢测方法。首先采用多通道灰度化预处理图像,并分别建立相应的背景模型,接着衡量新像素点和背景模型的相似性,分类出前景和背景像素,然后运用形态学操作和区域标记法滤除噪声,检测出运动目标,并通过区域标记定位目标。本文从不同角度开展对比实验,结果表明基于多通道背景提取算法的车辆检测能够有效提高车辆检测的精确性和完整性。
关键词:车辆检测;多通道;前景融合;区域标记法
中图法分类号 TP391.41 文献标识码: A 文章编号:1009-3044(2018)11-0203-04
Vehicle Detection Based on Multi-Channel Background Extraction Algorithm
ZHOU Sheng1,WANG Xiao2 ,WANG Zi-hao2 ,TAO Jing-yi2*,SHAO Ye-qin2,LI Ze-hui2 ,Hu Bin1 ,CHEN Jin-hua3,WANG Qiong4
(1.School of Computer Science and Technology, Nantong University, Nantong 226019, China;2.School of Transportation, Nantong University, Nantong 226019, China;3.Modern Education and Technology, Nantong University, Nantong 226019, China;4.School of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing 210094, China)
Abstract:Vehicle detection is a basic problem in intelligent transportation. To effectively detect vehicles, this paper proposes a vehicle detection method based on multi-channel background extraction algorithm. Firstly, the multi-channel gray-scale preprocessing images are used to establish the corresponding background model.Secondly, the similarity between the new pixel and the background model is measured, classifying the foreground and the background pixels. Finally,our method filters the noise by the morphological and regional markers, detects the moving target, and locates the target through the area marker. In this paper, extensive experiments are carried out. The experimental results show that vehicle detection based on multi-channel background extraction algorithm can effectively improve the accuracy and completeness of vehicle detection.
Key words: Vehicle detection; multi-channel; Foreground fusion; area marking method
随着社会经济的快速发展,道路上汽车的数量不断增加。机动车肇事逃逸、机动车逆行、机动车违停等交通违法行为显著增加。为了杜绝上述违法行为的发生,公安部门在各个路口和关键地点安装监控,对车辆的异常行为进行记录。目前,大部分网点还是采用人工方式检查视频中车辆的异常行为。因此,需要一种能自动从视频中检测目标(车辆)的方法,并在此基础上进行车辆跟踪和异常识别,最终检测出车辆的违章行为。
在复杂环境下,有诸多因素影响着运动目标检测的效果和质量,这给运动目标检测带来了很大的挑战,也使复杂背景下的运动目标检测成为研究热点[1-5]。目前常用的运动目标检测方法有光流法[6],帧间差分法[7],背景减除法[8,9]等。……
