Segmentation And Tracking Of The Object Based On The Active Contour Model

The research content of the thesis is an important part of the project"Remote display and remote control of intelligent robot" which is belong to"211 engineering" project of academic building of Harbin Engineering University.The motion analysis to a moving object in sequential images is an important part in computer vision technology. The motion analysis consists of object capturing (determination of the initial tracking position), object recognition (or segmentation), object tracking, and object understanding. The object capturing, segmentation and tracking provide the foundation for motion analysis. And a well-segmented and tracked result will lead into a correct object understanding. In order to allow computer vision system automatically segmenting and tracking the object continuously and at the meantime automatically gathering the object contour information, the research has been concentrated on the techniques of moving object segmentation and tracking based on active contour model in sequential images. In this thesis the concept of the active contour model and the potential of the associated improvement are introduced and then the process of determining the object initial tracking position and the process of the object tracking itself are further discussed.While determining the object initial position the algorithm is built upon polygon shaped active contour model for the object segmentation in an image. And also the image differential technique, Cellular Neural Network (CNN), and SUSAN (Smallest Univalue Segment Assimilating Nucleus) algorithm are used for the initialisation of the contour curve and its parameter settings of the active contour model. The position of the moving object is first determined by applying inter-frame differential gradient algorithm. When the region of interest (ROI) has been established, SUSAN algorithm is applied for obtaining the object contour...