一种基于背景预测的红外杂波抑制新方法

针对复杂背景下的红外弱小目标的检测问题,提出了一种中值二维最小均四阶矩(MTDLMF)自适应红外图像背景杂波抑制算法。该算法将观测数据经过一组低阶的自适应子滤波器,每个子滤波器的输出作为输入数据的一个预测,将所有的子滤波器的输出通过一个非线性中值滤波器,其输出作为最后的预测结果。该算法可以降低大量由于强边缘而引起的虚警,能够更有效地抑制杂波并且保留信号。实验结果显示,该算法对复杂背景下红外弱目标图像背景杂波抑制具有良好效果。

Novel method for IR image clutter suppression based on background prediction

To investigate the image character of the dim small IR target detection,a adaptive method based on median two dimensional least Mean fourth(MTDLMF) for IR image clutter suppression is proposed.This method firstly filters the observed image by a group of low order adaptive sub-filters,the outputs of which are the prediction of the input image.Then processes all the outputs of sub-filters by a nonlinear median filter,the output of which is the final prediction value.A mass of false-alarm resulted by strong edge can be depressed by the method,meanwhile clutter can be effectively suppressed and signal is preserved mostly.The experimental results show that this method is effective for clutter suppression of IR dim small target image in challenging condition.