【摘要】 针对高速公路智能交通系统中的复杂场景变化、阴影和车辆遮挡等影响车辆检测率的问题,提出了使用基于高斯混合模型的自适应方法来建立和更新背景,采用基于HSV的邻域均值快速阴影消除算法提高阴影消除的速度。对于遮挡车辆,采用基于Kalman滤波的识别算法进行遮挡车辆的识别,然后使用基于模板匹配的金字塔分级搜索算法对遮挡车辆进行细分割。实验结果表明,该算法既简单又有效,车辆的检测率达到97%,完全满足车辆检测的要求。更多还原

【Abstract】 Aiming at the problem that complex scene changes,shadows and overlapped vehicle affect the detection rate of vehicles in intelligent transportation system for the highway,an adaptive method based on Gaussian mixture model to build and update the background is proposed,a fast algorithm of local-mean based on HSV model for shadow elimination is used for improve the performance of shadow elimination.For the overlapped vehicle,first the overlapped vehicle identification algorithm based on Kalman filtering is adopted for overlapped vehicle identification,and then the search algorithm of hierarchical pyramid based on template matching is used for overlapped vehicle segmentation.Experimental results show that the used algorithm is simple and effective,the vehicle detection rate is over 97%,and it can fully meet the requirements of vehicle detection.更多还原