时域递归滤波算法中k系数确定方法的研究

摘要

Time-recursive low-pass filtering which belongs to nonlinear adaptive filtering is the main method in real-time for detecting small target with condition of complex backgrounds. The key of filtering is to determine the control parameter k. Actually, several proposals lie in the derivation of the parameter k. The value of this parameter must be a function of the prediction error, therefore it is able to correct filtering algorithm based on target motion. Comparison in hardware application between these proposals is discussed and oriented.%时域递归低通滤波是高精度光电实时跟踪控制系统在复杂背景条件下提高信噪比，检测运动弱小目标的主要方法，属于自适应的非线性滤波。这种非线性滤波的关键在于实时确定滤波算法中的k系数。这个k系数是预测偏差的函数，它能够根据目标运动的情况实时校正滤波算法。通常在确定这个系数时有不同的算法，通过比较这些算法在硬件上实时实现的差别，确定出适用于光电实时跟踪控制系统比较理想的时域递归低通滤波方法。