基于均值反差滤波的红外小目标检测算法

A Algorithm of Infrared Small Targets Detection Based on Average Absolute Difference

在已有背景预测算法的基础上,提出了一种基于均值反差法作为不相似度量对图像进行滤波,在增强小目标的同时抑制了复杂背景,并通过自适应的背景预测算法,实现了对小目标的快速检测。实验结果表明,该算法对小目标适应性强,能大幅度地提高目标/背景的信噪比,从而快速而准确地检测出小目标的位置。

Based on background forecast algorithms existed already, an average absolute difference filter algorithm is presented as the dissimilarity measurement between object and background region. The system utilizes this filter to enhance small target and suppress the clutter, and then it forecasts self-adaptively background. Finally, the system segments and obtains the small targets by the setting threshold. The experiment results indicate that the algorithm has excellent adaptability to small object shape and greatly improved SNR performance.