针对因复杂背景导致低信噪比的弱点目标探测率降低的问题,首先分析了从红外图像中探测弱点目标时,由于复杂和缓变背景下潜在目标探测率不同,而导致目标探测率降低的理论依据;并在该分析的基础上,提出了一种基于背景自适应调整的红外点目标探测算法。该方法利用鲁宾逊(Robinson)保护滤波器从经过预处理的图像中提取潜在目标;通过复杂背景模糊隶属度函数将图像映射到模糊特征平面,并由该特征平面计算背景调整因子,以对提取的潜在目标进行加权调整,从而降低了复杂背景的影响。实验结果表明,该算法可以显著提高复杂背景下红外点目标的检测概率,并且能够探测出信噪比为1的目标。

In order to advance the detection ratio of dim target in clutter background, the reason of low target detected radio from infrared image that contain both complex and flat background is analysed. And based on this analysis, an algorithm of dim target detection based on adaptive background adjusting is advanced. This algorithm uses Robinson guard filter to distill potential target point. And clutter image fuzzy membership function is defined to map the image to fuzzy plane. The background adjusting coefficient is computed by this fuzzy plane to modulate the potential target point. So the influence of the clutter background of the image is reduced. The experimental results have proved that this algorithm can obviously improve the probability of the point target detection, and it can detect dim targets within infrared image with signal-to-noise ratio higher than 1.