基于视频图像的车辆计数新方法

A New Method for Vehicle Counting Based on Video Images

针对道路车流量检测问题,从便捷性、实时性角度出发,结合视频图像处理技术,对视频车辆计数进行了研究。直接在RGB图像中进行自适应背景更新,以此为基础,对RGB图像进行背景差分,提取出运动车辆区域,避免了复杂环境下图像灰度化过程中的信息丢失;利用当前帧和背景帧的HSI颜色空间信息来滤除阴影;通过在视频图像中设置固定虚拟检测区,实现对车辆的计数。实验结果表明,该方法计算量较小,白天情况下的计数准确率在89.58%以上;夜间的计数准确率较低,还需进一步研究改进。

A new vehicle counting method is put forward based on video images processing technology for road traffic statistics considering the reliability and the real-time. Adaptive background updating was done directly on RGB images,based on which,vehicle movement area could be extracted via background subtraction. This could avoid the loss of information in the course of gray in complex environments. And then shadows were filtered using the H component of HSI color model of current frame and background frame. The final counting work was done via setting a fixed virtual detection zone. Experimental results show that vehicle counting accuracy based on the method proposed by this paper is more than 97. 22% during the daytime. In addition,the computing time is less. The accuracy of the count is low at night,which is needed to be improved.