飞行小目标自适应双波门跟踪算法

Adaptive double gate tracking algorithm for weak aerial target

波门跟踪是一种利用实际图像处理的区域范围的视觉跟踪算法。在飞行小目标信噪比较低的情况下,现有方法往往不能获得稳定的跟踪结果。针对小目标跟踪的特点,提出了一种自适应双波门的跟踪算法。该算法采用帧间差分算法检测小目标,通过质心标定修正当前帧目标的初始位置,并利用最小二乘估计预测算法实现目标范围的估计,从而实现自适应跟踪的目的。实验结果表明,即使在飞行小目标被短时遮挡情况下,利用本算法也能稳定地检测出它的位置,能根据目标的轮廓自适应地确定目标波门形状,并能有效解决目标跟踪偏移和短时丢失等问题。

The gate track is a visual tracking algorithm using actual image region to process.While the SNR of weak aerial target（WAT） is lower,the existing algorithms is so finite that we can not track stability.According to the WAT tracking feature,this paper proposes an algorithm of adaptive double gate tracking.This algorithm has certain robustness to the target migration and shot missing,.Using this method,the initial position of WAT centroid calibration can be revised,the scope of WAT estimated by using the least square estimation,and the tracking realized finally. The experiment result indicates that this method can not only detect the WAT precisely even if it is missing, but also determines the gateg shape adaptively from its contour.