帧差法和Mean shift算法融合的高速无人机目标跟踪

Fusion Algorithm of Frame Difference and Mean Shift for High-Speed UAVs' Tracking

为实时跟踪高速飞行无人机,图像跟踪算法必须满足快速性和准确性要求。文章给出一个融合算法,将帧差法和Mean shift算法的优势结合起来。2个算法平行运行,差帧法实现快速跟踪,Mean shift算法则用于对帧差法结果进行准确度修正。还利用Kalman滤波技术对计算周期内的无人机运动位移进行补偿,进一步提高实时跟踪的准确性,并给出Matlab仿真例子验证本文方法的有效性。

In order to track high-speed UAVs in real time, the visual tracking algorithms need to satisfy the requirements of accuracy and rapidity. In this paper, a fusion algorithm was presented that combined the advantages of two typical algo- rithms, frame difference and Mean shift. The two algorithms were run simultaneously the frame difference was used for rap- id tracking, while Mean shift was used for improving the precision of the tracking of frame difference. To get more precise tracking, the Kalman filter was used to compensate the displacement of UAVs moving during the implement of the tracking algorithm. The effectiveness of fusion algorithm was illustrated by a simulation example in Matlab environment.