论文部分内容阅读
滤波滑窗的选取是SAR图像相干斑滤波的关键问题之一,从SAR图像中均匀区域的幅度和空间特性出发,提出了一种基于置信区间与形态学重构的自适应滑窗方法。首先在初始滑窗内进行强度筛选,依据初筛结果的异质程度自适应调整置信区间,得到在幅度上连续的初始自适应窗;接着以初始自适应窗为掩膜,通过空间邻接约束进行形态学重构,进一步得到在空间上连续的、形状任意的自适应滑窗。实验结果表明,与固定窗法和改进的Sigma方法相比,该方法能够获得更准确的滤波滑窗。
The selection of filter sliding window is one of the key problems of SAR image coherent point filtering. Based on the amplitude and spatial characteristics of uniform region in SAR image, an adaptive sliding window method based on confidence interval and morphological reconstruction is proposed. Firstly, intensity screening is carried out in the initial sliding window, and the confidence interval is adaptively adjusted according to the heterogeneity of the preliminary screening results to obtain an initial adaptive window in amplitude. Then, using the initial adaptive window as a mask, the spatial adjacent constraint Morphological reconstruction, and further to be spatially continuous, the shape of any adaptive sliding window. The experimental results show that this method can obtain more accurate filtered sliding window than the fixed window method and the improved Sigma method.