改进混合高斯模型的运动目标检测与跟踪算法

Tracking and Detection Algorithm of Motion Targets Based on Improved Gaussian Mixture Model

针对混合高斯模型的运动前景更新难题,为了提高运动目标跟踪精度,提出一种改进高斯混合模型的目标检测与跟踪算法;首先提取目标特征建立目标分类器,并将目标从前景标记出来;然后通过多目标跟踪将目标分为多种运动模式;最后采用高斯混合模型对目标进行跟踪与分类,并采用仿真实验测试算法的性能;结果表明,文章算法不仅提高了目标检测与跟踪精度,而且可以满足目标跟踪的实时性要求。

Gaussian mixture model can not solve the problem of updating motion background, in order to improve the tracking precision of moving targets, a novel tracking and detection of motion targets based on improved Gaussian mixture model is proposed in this paper. Firstly, target features of target are used to build classifier, and the target from the foreground is marked; secondly the target is divided into for multi--motion modes by using multi target tracking method; finally, the results of tracking and classification were fused to obtain the final goal location by Gauss mixture model. The results show that the proposed algorithm not only has improved target detection and tracking accuracy and can better meet real--time requirements of the target tracking.