利用Gamma CFAR进行SAR图像目标检测

Modified target detector for SAR image by quadratic Gamma kernels

利用二项式Gamma核进行SAR图像目标检测,并提出自适应选取扫描窗大小的方法,给出了一种基于改进的二项式Gamma核SAR图像检测方法提高二项式Gamma核检测器的性能。用于高分辨率单视单极化的极不均匀的SAR图像,能够在较小虚警率下将人造目标全部检测出来,而且能很好地保持人造目标的结构信息。实验证明所提方法的有效性。

CFAR detector is widely used as a target detector in SAR because of its simplicity and effectiveness. However, when used in an extremely heterogeneous SAR image, it leads to a bad performance. So, a new target detector for SAR image by quadratic Gamma kernels with adaptive guard-band size is presented. This method is tested with high resolution single look single polarity SAR data. The results show that it can detect all the artificial targets with lower false alarm rate, and in the same time the detected targets preserve their structure information very well.