基于改进高斯混合模型的运动目标检测与跟踪

摘要

In order to improve the accuracy and reliability of moving object detection and tracking,a moving object detec-tion and tracking method based on improved Gaussian mixture model is proposed. The improved Gaussian mixture background model is established to segment the moving target image into blocks. The continuity of the neighborhood frames is utilized to up-date the parameters of the moving target image,and then the whole moving target is extracted for segmentation. The given pixel of current frame is matched with the target image to reduce the distribution quantity and calculated amount of the Gaussian mix-ture model. According to the size,shape and color information of the moving target after block processing,the global matching for moving target is conducted to realize the real-time detection and tracking of moving object. The experimental results show that,in comparison with the current moving target detection and tracking methods based on Gaussian mixture model,the pro-posed method has simpler calculation process,faster detection speed,and more reliable detection results.%为了提高运动目标检测与跟踪的精确性与可靠性,提出一种基于改进高斯混合模型的运动目标检测与跟踪方法.首先,建立改进高斯混合背景模型,对运动目标图像进行分块处理,利用相连帧的连续性对运动目标图像的参数更新,提取完整的运动目标并进行分割;其次,将给定的当前帧像素点与目标图像进行匹配,减少高斯混合模型的分布数量和计算量,根据分块处理后的运动目标的大小、形状以及颜色信息完成运动目标全局匹配,实现运动目标的实时检测与跟踪.实验结果表明,与目前的高斯混合模型对运动目标检测与跟踪的方法相比,所提方法计算过程较为简单,具有更快的检测速度和更可靠的检测结果.