Active Target Tracking Based On Visual Platform Of Unmanned Aerial Vehicle

Rotorcraft unmanned aerial vehicle(UAV)is a type of aircraft which can rise vertically,hover at a fixed point,and fly at low speed.A rotorcraft UAV equipped with visual platform can be widely used in the military reconnaissance,disaster relief,road monitoring,industrial transport and other service fields.In this paper,we study the active tracking technology based on visual platform of UAV.A novel tracking algorithm for moving targets which is suitable for this moving platform is developed.The system realizes tracking ground moving targets actively.First of all,we propose a system's framework including hardware and software after reviewed relevant literatures.According to the requirements of active target tracking system of the vision platform on an UAV,the hardware composition and software architecture are analyzed and designed based on the principle of system reliability,mobility and accuracy,so that it can be effectively used in the active target tracking task.The core of object tracking system is a robust tracking algorithm.In order to solve tracking failures caused by objects deformation,occlusion and fast motion,a novel algorithm called MS-TLD which under the Tracking-Learning-Detection framework is proposed.The algorithm reconstructs a new tracker with the scale-adaptive mean-shift method(ASMS).By introducing color histogram features and scale-adaptive,the new tracker can track objects with deformation and fast moving.We establish a new tracking-detection feedback strategy—the inter-correction between tracker and detector.Therefore,the new algorithm has better robustness when objects are occluded.We use OTB-50 standard dataset to verify and evaluate our method.The experimental results show that the proposed algorithm can overcome the tracking failures caused by objects with deformation,occlusion,fast motion,as well as background clutters,and has better tracking accuracy and robustness compared with TLD and other three classic algorithms.Finally,we develop a ground station control program based on MFC and OpenCV library.Through the analysis of the camera coordinate system,the image coordinate system and the pixel coordinate system,we transform the target real coordinate to the pixel coordinate.Our visual platform system that equipped on a DJI Inspire1 Pro series UAV can active track a moving-object-successfully.