A Moving Target Tracking Based On Hausdorff Distance Algorithm

Visual analysis of moving object's motion is one of the most important research topics in the domain of computer vision. It is also an active filed which has interested many researchers in recent years. Moving object's motion analysis aims at attempting to detect, identify and track object from image sequences, and more generally, to understand and describe their behaviors. Motion detection and tracking belong to the low-level vision problems, and behavior understanding and description belong to the high-level vision problems.This paper aims at the low-level vision problems including object motion detection and tracking. We investigated the approaches of how to model the interesting object in video sequences and track the contour in the following frames,by which we can analyses the moving mode and track it. That is ,at first we use GVF Snake model to extract the interesting object contour in the first frame,which was matched as template in following frames.The matching algorithm we used is maximum likelihood hausdorff distance,which used the likelihood replace the distance of image characteristic pixel and model pixel,and a Multi-resolution searching strategy was used to search the best solution.The primary work in the thesis as follows:1. A new hausdorff algorithm was used ,which is different with average hausdorff algorithm .It uses likelihood degree to replace the distance of model points and image points,so statistics information and prior knowledge can be used. 2. Multi-Resolution searching strategy was used to search the best matching position,whose advantage is that it can find the undesired solution and prune it quickly,and enhance the whole system performance.Experimental results show that it can reduce the searching time when it get the same searching precision with other searching strategy.