基于非参数统计的综合孔径图像弱小目标检测

Puny target detection in synthetic aperture radiometer image based on nonparametric statistics

针对综合孔径微波辐射图像中目标和复杂背景的边缘振荡对目标检测带来的影响,提出一种基于非参数统计的弱小目标检测方法.在对综合孔径辐射图像特征进行研究的基础上,采用非参数统计中的秩统计量描述图像中目标及复杂背景边缘的过渡纹理,给出综合孔径辐射图像的非参数统计描述模型,利用目标及复杂背景边缘各方向秩相关系数不同的特征,将目标与复杂背景有效地分离,实现对复杂背景的抑制和目标检测.仿真及实验结果表明该方法能有效地对综合孔径辐射图像中的弱小目标进行检测.

To solve the impact from the oscillation at the edge of the target and background in synthetic aperture microwave radiometer image,apuny target detection method based on nonparametric statistics was presented.The characteristics of synthetic aperture radiometer image was researched,the rank analysis method was introduced to describe the edge transition texture,the nonparametric statistics model of synthetic aperture radiometer image was built,and the target and background was separated by the different characteristics of their rank correlation between different directions.The simulation and experimental results show that the method can efficiently detect the puny target in synthetic aperture radiometer image.