中文摘要：

针对交通场景中车辆距离过近和相互遮挡的问题,提出了利用车辆边缘的遮挡车辆曲线分割算法。首先在图像分块的基础上检测出车辆区域,根据车辆区域的长宽比和占空比进行多车判断;然后对车辆区域进行HSI空间亮度均衡化和平滑处理,并利用一维最大熵法分割图像;最后确定车辆区域的中心线,并提取车辆的边缘轮廓,曲线分割遮挡车辆。实验结果表明,算法能够按照车辆的边缘轮廓准确分割遮挡车辆,与其他算法相比,在满足实时性的前提下具有较高的识别率,查全率和查准率均可达到90%左右。

英文摘要：

When using the video surveillance system to detect the traffic scene,the false detection often existsed when many vehicles were close to each other,and presented a segmentation algorithm based on contour curve of occluded vehicle. Firstly,it divided the image into blocks and detects the vehicle regions,and considered the width /height ratio and occupancy ratio to make detection judgment. Then,it used the methods of HIS histogram equalization and image smoothing to process the image,and used the one-dimensional maximum entropy method to split the image. At last,determined the center line of the vehicle,and identify edges of vehicles,which were regarded as segmentation curves,then split the occluded vehicles. Experimental results demonstrate that,the proposed method has a high recognition rate and can segregate the overlapped ones accurately and completely. Compared with the other methods,it has a high segmentation result and higher recognition success rate,and the recall and precision can reach 90%.