论文部分内容阅读
传统的全局检测算法假设目标占整个背景中较小的一部分,将SAR图像中的所有像素用于估计杂波概率密度函数,容易造成检测阈值的增大从而对不太明显的SAR图像舰船目标产生漏检。提出了一种新型自适应背景窗的检测算法,该算法通过设置一种新型的窗口对潜在的目标予以剔除,鉴于Parzen窗对杂波背景的良好的估计性能,对剔除后的杂波背景采用Parzen窗进行非参数化的杂波模型估计,进而确定检测阈值,完成目标的检测。相比全局检测算法,提出的SAR图像舰船目标检测算法减少了漏检数量,检测性能得到了改善。实测SAR图像的检测结果表明了该方法的有效性。
The traditional global detection algorithm assumes that the target occupies a small part of the whole background and all the pixels in the SAR image are used to estimate the clutter probability density function, which easily leads to the increase of the detection threshold and thus to the less obvious SAR image ship target Produce missed inspection. A new adaptive background detection algorithm is proposed. The algorithm removes a potential target by setting a new type of window. In view of the good performance of the Parzen window for the clutter background, the clutter background is removed Parzen window to non-parametric clutter model estimation, and then determine the detection threshold to complete the target detection. Compared with the global detection algorithm, the proposed SAR target detection algorithm reduces the number of missing detections and improves the detection performance. The test results of measured SAR images show the effectiveness of this method.