基于视频的目标检测与跟踪算法研究计算机应用技术专业论文.docxVIP

Video-based target detection and tracking algorithm ABSTRACT Target detection and tracking is one of the main content in the research of the digital image processing and pattern recognition and computer vision field. It has a wide range of applications in the field of robot navigation, security and surveillance systems, military defense, and virtual reality. The main work in this paper is researching of moving target detection and moving target tracking. The main contents are as follows: It proposes a kind of moving target detection algorithm with combining of two-way differential phase multiplication and background subtraction in target detection. First of all, it does the two-way frame difference in the video images. Secondly, it does the Binary segmentation to the differential image with the OTSU threshold method. Before the segmentation, it has to judge the value of Binary threshold obtained from the split each frame. If the value is too small, it recognizes there is no foreground object in this frame and it doesn’t need to do the target detection. If there is the foreground object, it has to obtain the foreground object area with the background subtraction and put it to be multiplied by detected target edge information derived from two-way differential phase multiplication to get the complete target area. And then, of the target binary image morphological processing, a complete moving target can be obtained. Thirdly, detecting of moving targets in the video sequence using two scanning method. Fourthly, the validity of the method can be proved through the comparison of experimental results. At first, it selects the particle filter algorithm with high robustness as the basic tracking algorithm in target tracking. Then, because of the poor tracking performance in the case of the situation that the foreground and background colors are similar or the target color information is not obvious for tracking algorithm based on the characteristics of a single color part