Moving Targets Tracking And Extraction Based On Transmission Lines

Along with the rapid development of the information technology and the growing of computer processing power and the continuous improvement of the of operation speed and the significant reducement of storage costs, moving target tracking technology has gradually become an important topic in the field of computer vision. Computer vision tracking technology involves a wide range of fields, including military target detection, robot visual navigation, video surveillance, human-computer interaction, intelligent transportation, etc. And the technology has been widely used in many areas.With the continuous development of the inspection robot technology based on transmission lines, the technology has improved steadily.the manual inspection has been partially replaced by robot inspection technology in some developed countries. Currently the robot inspection technology has become one of the important development direction in the environment of overhead transmission lines. It will be a big trend that the robot inspection technology gradually replaces manual inspection. Moving targets are likely to harm the safe operation of transmission lines in overhead transmission line environment, but the work of these detection and tracking of moving targets will be completed by computers. So, it has an important significance for the study of moving targets detection and tracking in the transmission line environment. The following work is mainly done for the topic in this article:Fist,an improved three-frame difference method is proposed in this paper.The algorithm is suitable for slower moving or smaller volume moving target de tection in overhead transmission line environment. And the algorithm feasibility is ex perimentally validated.Finally, a multi-frame-differencing-method is proposed with th e superposition of improvement three-frame-differencing-method principle.Then,an improved target tracking algorithm based on the right values is presented in this paper. mean shift algorithm based on probability distribution graphs and kalman filter algorithm combin in the algorithm. The color characteristic and the centroid feature are used in the algorithm.The algorithm enhanced target tracking accuracy.The final,An adaptive object tracking algorithm is presented in this paper. The algorithm enhances the flexibility of the target tracking algorithm based on the right values and can be used for the target scene of high precision tracking.