Similar targets detection and location indoors by UAV

With the increasing demand of UAVs application in target reconnaissance, it is a more complex task to identify and locate similar targets. The small difference between similar targets and the limited computing power of UAVs on board mean that the perception algorithm needs to have high target differentiation and fast accuracy. In order to solve the above problems, this paper proposes a similar target detection and target absolute position localization algorithm based on yolo-v3 partial detection, including a detection module and a positioning module. The detection module uses the speed of yolo-v3 and the effectiveness of small target detection, and improves by adding geometric constraints to achieve similar target differentiation. The positioning module obtains the accurate and absolute position of the target for the UAV platform. The actual flight experiment is carried out and verify that the proposed algorithm works well in real time under the UAV platform.