复杂场景下由背景物体产生的虚警是红外小目标检测中的难点, 在目标信号微弱时问题尤为突出。为了抑制虚警, 提出一种基于局部对比度特征的小目标检测算法。首先, 利用一种新型 Top Hat变换检测出图像中的潜在目标区域; 然后分析潜在目标区域及其邻域特性, 计算本文提出的局部对比度特征; 最后基于此特征提取目标区域。对合成图像和实际红外图像的实验表明, 本算法能有效地抑制背景景物造成的虚警, 取得很高的检测率, 并且对目标强度的变化以及目标所在背景的变化均具有一定的适应性。

False alarms generated by complex scenes of the background are a challenging problem for infrared small target detection, especially when the target is dim. To reduce false alarms, a small target detection algorithm based on local contrast is proposed. First, a modified top hat transform is applied to detect regions that are likely to contain a target. Then for each of the regions, we calculate the proposed local contrast measure. The target regions are then detected based on such measure. Experimental results on synthetic and real data show that the proposed method reduces false alarms generated by background and achieves high detection accuracy. Moreover, the algorithm is shown to be robust to the variation of the target intensity and fluctuation of the target neighborhood.