为了更有效地检测出红外弱小目标,通过分析红外图像中弱小目标与其邻域背景的特征差异性,提出了一种基于统计特征和桥梁方法的红外小目标检测算法。在滑动窗口范围内提取像素值的均值、方差等特征,根据这些统计特征和桥梁方法判断该窗口范围内有无红外小目标;如果存在小目标,记录下其位置;对小目标区域进行二次筛选。研究结果表明,所提算法相对于较经典算法,虚警率降低了58%以上。

In order to detect dim the small infrared targets more effectively, a small infrared target detection algorithm based on statistical features and bridge method is proposed by analyzing the difference between the small dim infrared targets in infrared images and their neighborhood backgrounds. The mean value, variance and other characteristics of pixel values are extracted in the sliding window range. According to these statistical characteristics and the bridge method, whether there are small infrared targets in the window range is determined. If there is a small target, its location is recorded, and then the small target area is screened twice. The research results show that the false alarm rate of the proposed algorithm is 58% lower than that of the classical algorithm.