Tthe Research On Methods Of Moving Target Detection And Tracking Based-on Ffrog Vision Characteristics

Detection and tracking of the moving target has been an important research topic in the field of computer vision. Many researchers at home and abroad have done deep researches on it and got rich results both in theory and technology. Meanwhile, as it is widely used in military and civil fields, the higher demand is proposed in moving target analysis especially for real-time, accurate and robustness under complex conditions. For traditional theory is showing insufficient to solve the related problems, the bionic vision based on biology is a potential solution. Therefore, this paper makes series of studies on the detection and tracking of the moving target mainly through analysis of the frog visual characteristics.First, the development situation on technology of detection and tracking of the moving target is introduced. The defects and difficulties of traditional theory in solving the related problems are analyzed, and an idea is put forward to learn the frogs which have sports vision expertise on visual characteristics and cognitive mechanism. It’s also explored that thinking of a new method is used for cognitive computing on detection and tracking of the moving target. It describes the frog visual cognitive characteristics and neural mechanisms which provide a basis for visual bionic.Then, the principle that the frogs can quickly and accurately detect moving targets is analyzed. Then, a kind of moving object detection algorithm is designed by simulating frogs’two kinds of ganglion cells in the retina and combining with the video image processing algorithm. The algorithm fuses the movement characteristics and edge characteristics in multi-angle and multi-level ways, uses a double threshold fusion method, and finally accurately detects the movement edge of target. On the one hand, it is based on the frogs’visual characteristics during the process of algorithm design. On the other hand, it makes reasonable use of the advantages of the video image processing methods.Finally, in connection with the scale changed of target in the process of tracking, an improved Mean-shift moving target tracking algorithm is proposed based on that the frogs’eyes is sensitive to the moving target and some related theories of tracking box and target scale. The algorithm is mainly divided into two parts:one is to build a movement-color histogram to describe the target by compromising the motion feature and color feature of the moving object; the other is that the size of tracking window is updated adaptively according to the object scale based on frogs’eyes feature and Mean-shift calculation. At the same time, the feature template is updated. Simulation results show that the method can locate moving target when scale changes.