Research On Video Tracking For Formation Flight Of Unmanned Aerial Vehicles

With a large number of applications of UAV in military and civilian aspects, the development ofUAV has been attached great importance of all countries. Due to single UAV not to complete task andtasks of higher complexity, UAV formation flight can cooperate to complete more complex tasks, andis more efficient. Therefore, the development of UAV formation flight is an important trend.Visual tracking technology can obtain the state parameters of interested target from videosequences, so in the formation flight visual information can be used for target detection and tracking,to determine information of direction and distance of neighbor UAV and to provide measurement forformation control.The design of target tracking system in this paper is divided into software algorithm design andhardware design. The software algorithm has target detection and target tracking. Target detection isthe foundation of target tracking. In the stage of target detection, the system firstly can obtain UAVsports images from visual sensor, and detect target from background after the image pre-processing,histogram, threshold segmentation and so on. Target tracking can obtain the location of the object.Tracking algorithm was KLT algorithm, and was used to track target detected from video image andcalculate the relative displacement of UAV in the image.According to the requirements of tracking system, in this paper the design of the hardwareplatform was based on the dedicated video processing DSP chip DM642, which included imageacquisition, image storage, image processing, image output, communications and pan movementcontrol. Video-driven, target detection, target tracking and pan control was realized based on CCS ofTI DSP, and physical simulation verification was developed based on DSP.The verification results show that the results of tracking is of higher accuracy, and can satisfyaccuracy requirements of application, so the system can be used for stable tracking of UAV formation,and provide information to solve UAV relative positioning.