基于形态学及小面核滤波的红外小目标检测方法

Infrared Small Target Detection Method Based on Morphology and Facet Kernel Filtering

鲁棒性和有效性是红外小目标检测的关键,小目标在复杂背景下往往被杂波淹没,存在较难检出的问题.基于此,提出了一种基于形态学及小面核滤波的红外小目标检测方法.该方法受到HVS人视觉显著性特征的启发,通过选取合适的结构元素进行形态学预处理,设计小面滤波核,在图像预处理的基础上进行卷积滤波,生成核滤波图,最后利用阈值进行分割检出小目标.实验结果证明,该方法增强了小目标的同时抑制了复杂背景杂波,此外,该方法在不同噪声及复杂杂波下均存在较高的鲁棒性和有效性.

Robustness and effectiveness are the key to infrared small target detection.Small targets are often submerged by clutter under complex background and are difficult to detect.An infrared small target detection method based on morphology and facet kernel filtering is proposed in this paper.This method is inspired by the salient features of the HVS human visual system,and selects the appropriate structural elements for morphological preprocessing.By designing a facet-based filter kernel,a convolution filter is carried out on the basis of the preprocessed image to generate a facet-based kernel filtering map.Finally,a threshold is used to segment the small target.The experimental results show that the proposed method can enhance the small target and suppress complex background clutter.In addition,the proposed method has high robustness and effectiveness under different noise and complex clutter.