时空域双边滤波背景抑制算法研究

摘要

The strong edge in an infrared image is an important factor restricting the infrared dim target detection probability and also an important source of target false alarm.By taking advantages of bilateral filter's characteristics of the Euclidean distance of neighbor pixel and the similarity of pixel gray value, as well as the slow-changing inter-frame background, we defined the spatial-temporal bilateral filter and verified the inhibitory effect of this algorithm in the good complex backgrond.Results show that it can retain the target energy substantial and improve the local signal-to-clutter ratio(SCR), the signal-to-noise ratio gain(GSNR) is greater than 3.6 and the target energy suppress factor is less than 0.2.%红外图像中的强边缘一直是制约红外弱小目标检测概率的重要因素,同时是产生目标虚警的重要来源.采用能够同时利用邻域空间欧式距离和灰度值相似性的双边滤波,并针对天基红外系统中帧间背景缓慢变化的特点,设计一种基于时空域的双边滤波法,实验验证该方法对复杂背景的强边缘具有很好地抑制效果,信杂比增益大于3.6,同时能够很好地保留目标能量,目标能量损失因子D-factor小于0.2,有效地提高目标的局域信杂比.