基于直方图形态的SAR图像灰度变换性能评估

以合成孔径雷达(syntheticapertureradar,SAR)图像灰度变换前后直方图形态一致性为依据,对现有的7种变换算法性能进行了评价。首先对原始图像进行灰度线性变换,以线性变换图像作为参考图像,对7种变换算法的结果图像进行评估。提出3个检测直方图形态差异的统计参数:相对均值偏量、相对方差偏量和相对偏态差。经真实SAR图像检验,表明基于辐射度统计多分辨率分析(RMA)的灰度变换方法性能最优。

Evaluation for the gray-scale transformation of SAR images by the shape of histogram

Based on the shape difference of histograms before and after gray-scale transformation of synthetic aperture radar (SAR) images, seven gray transform methods are evaluate. Firstly linear gray-scale transformation for the original image is made. Then the transformed image is taken as a reference to evaluate the resulting images by seven gray transform methods. Three statistics to measure the shape difference of histograsm are proposes: relative mean error, relative variance error and relative skewness error. Test on the actual SAR images shows that the gray transform algorithm based on radiometric multiresolution analysis (RMA) is the best.