Using shape feature matching to track moving objects in image sequences

The paper develops a methodology to track moving objects in image sequences. The methodologies in tracking moving objects are influenced by the natural environment and motion characteristics. This paper introduces an algorithm to track moving objects under a natural environment (high noise ratio background). This algorithm can be applied from low noise ratio background to high noise ratio background. It also provides adjustable parameters to fit into different situations. It makes this algorithm flexible. The algorithm extracts moving objects by the difference between consecutive image frames and analyzes the shape features of moving objects. After shape analysis, the different moving objects are distinguished based on different shape feature characteristics including motion coordinate, speed, and direction. The original moving path can be found by coordinate transform. Some restrictions are applied. The moving object has to be a rigid body. The moving path has to be uni-directional. One camera is used which has to be stationary to take image sequences.