Unmanned aerial vehicles (UAVs) are prone to losing their targets when tracking moving objectives. A tracking strategy is proposed herein that enables the standoff tracking of a moving target using a vision system, which significantly reduces the occurrence of target loss. The strategy combines the Gimbal Control Algorithm based on Motion Compensation (GCAMC) with the Improved Reference Point Guidance Method (IRPGM). The GCAMC utilizes the attitude of the UAV and the deviation of the target from image center as the feedback. The target can be kept within the field-of-view (FOV) of the camera when the gimbal model is unknown. The IRPGM generates straight or circular paths according to the speed and potition of the target, while the UAV will continuously track the generated trajectory to achieve the objective of target tracking. To validate and demonstrate the tracking performance of the proposed strategy, a closed-loop visual simulation platform was devised and implemented to simulate the process of target tracking. The results of the simulation demonstrate that by using the proposed approach, the UAV can enter the desired trajectory quickly when its initial position and flight direction are arbitrary.

追踪运动目标时，无人机容易丢失目标。本文提出了一种跟踪策略，该跟踪策略允许使用视觉系统对运动目标进行对峙跟踪，从而大大减少了目标损失的发生。该策略将基于运动补偿的云台控制算法（GCAMC）与改进的参考点制导方法（IRPGM）相结合。GCAMC利用无人机的姿态和目标与图像中心的偏离作为反馈。万向节模型未知时，可以将目标保持在摄像机的视野（FOV）内。IRPGM根据目标的速度和强度生成直线或圆形路径，而无人机将连续跟踪生成的轨迹以实现目标跟踪的目的。为了验证和证明所提出策略的跟踪性能，设计并实施了一个闭环视觉仿真平台来模拟目标跟踪过程。仿真结果表明，通过使用本文提出的方法，当无人机的初始位置和飞行方向为任意时，无人机可以快速进入期望的轨迹。