基于改进运动历史图像的多运动目标实时跟踪

Real-time detecting and tracking of multiple moving object based on improved motion history images

提出了一种利用视频图像对运动目标进行实时检测与跟踪的新方法。该方法利用基于改进的时间片的运动历史图像(tMHI)的灰度阶梯轮廓方法对多个运动目标进行检测,通过卡尔曼滤波器对多目标进行跟踪,并得到了各个运动目标的轨迹曲线,进而实现了对视频图像中多目标的跟踪。同时,该方法对多个目标的遮挡问题获得了明显的改善效果。实验结果表明,该方法能够对复杂场景下的多个目标进行有效的识别和准确的跟踪,系统的实时性强,识别率高,而且该方法对于复杂视频监视系统场景中的光照变化、雨雾等干扰具有较强的稳健性。

A new method for segmentation and tracking real-time moving object from video image was proposed. By using the step-down grey value silhouette of timed motion history image （tMHI） to detect moving objects, and using Kalman filter to track multiple objects, this method gained trajectory of each object and achieved the tracking of multiple moving objects from video image, meanwhile, part conglutination was basically solved. The method was tested in a traffic video, some moving objects were segmented and tracked effectively. It appeared that the method has good locating robust for outdoors video surveillance system and fast process speed, meanwhile the method has good robustness for the change of ray and the disturbance of rain and fog.