中图法分类号：TP3914文献标识码：A文章编号：10089612014.DOC

中图法分类号：TP391.4 文献标识码：A 文章编号：1006-8961(2014) - - 论文引用格式： 多特征融合的车辆阴影消除 邱一川1，张亚英1，刘春梅1 1.同济大学嵌入式系统与服务计算教育部重点实验室，上海 200092 摘 要：目的 提出一种基于颜色特征和边缘特征相融合的算法，实现对复杂交通场景中车辆阴影的检测和消除。方法 首先，通过经典混合高斯背景建模方法帧差法获取运动目标前景。其次，利用边缘差分、形态学运算。为提高算法效率，对前景区域进行阴影评估，从而判断是否有必要进行阴影检测和消除。结果 实验结果表明，该算法能够有效消除车辆阴影，具有良好的准确性和鲁棒性。结论 此算法结合颜色和边缘两种特征，弥补基于单个特征方法的单一性，降低由于阴影区域边缘复杂、车辆颜色与阴影颜色相近等原因造成的阴影误检率，阴影消除效果良好。关键词 ：颜色特征；边缘特征；多特征融合；阴影评估；阴影检测；阴影消除 Vehicle shadow removal with multi-feature fusion Qiu Yichuan1, Zhang Yaying1, Liu Chunmei1 1. The Key Laboratory of Embedded System and Service Computing, Ministry of Education, Tongji University, Shanghai 200092, China Abstract：Objective A novel algorithm combining color feature and edge information is proposed to detect and remove vehicle shadow in complex traffic scene. Method Firstly, a background model is built by classical Gaussian Mixture Background Modeling method, and the moving vehicle foreground is obtained through frame difference. Secondly, a serial fusion strategy combining color feature and edge information is applied to detect and eliminate the vehicle shadow. On the basis of vehicle shadow detection by edge information method of the moving target foreground, RGB color feature detection method is implemented to further detect the shadow area and get a more precise result. To effectively detect and eliminate the shadow, edge difference and morphological processing methods are used during the operations. In order to improve the efficiency of the algorithm, shadow assessment is periodically evaluated on the foreground area to determine the necessity to apply the proposed algorithm. Result By comparison with SP, SNP, DNM1 and DNM2 algorithms, the proposed method realizes about 10% advance on shadow detection rate and shadow recognition rate. High accuracy and robustness of the proposed shadow removal method is revealed by the testing results, and the effectiveness of the method is also validated. Conclusion The proposed method combining colo