基于负指数变换和Mean-shift算法的SAR图像可视化方法

Visualization method for SAR images based on a negative exponential transformation and mean-shift algorithm

为了便于合成孔径雷达(SAR)图像的人工视觉解译,提出了一种SAR图像的可视化方法。用参数化的方法对SAR图像数据进行负指数变换;用Mean-shift算法估计变换后图像直方图分布的众数调整变换参数;对原SAR图像重新进行变换。实验结果表明:与传统方法相比,该方法得到的图像能适当的拉伸暗目标像素值,同时能较好地显示亮目标的细节;而且具备一定的鲁棒性,适用于不同场景的SAR图像。

A visualization method was developed for synthetic aperture radar（SAR） images to facilitate visual interpretation of SAR images.A negative exponential transformation is performed on the original SAR image with the transform parameter then modified based on the mode of the output image histogram which estimated by the mean-shift algorithm.The original image is then re-transformed using the modified parameter.Test results show that this method appropriately stretches the dark pixels while the bright details are preserved.The robustness of this method is demonstrated using SAR images of various scenes.