双树复小波域马尔科夫的遥感图像分割方法

针对多尺度高分辨率遥感图像像素分割在降噪时丢失大量高频信息及单一像素孤立性问题，提出了一种双树复小波变换(dual-G tree complex wavelet transform,DT-CWT)和模糊马尔科夫随机场(fuzzy markov random field,FMRF)模型相结合的无监督遥感图像分割算法。首先通过DT-CWT 遥感图像进行多尺度分解，并采用Bayesian阈值法对分解后的高频分量进行去噪，以增强图像的细节和边缘的表达能力并有效保留图像的主要高频信息；然后采用FMRF分割算法分别对重构后各层分量进行分割，以充分考虑像素分割的模糊性和像素邻域间的相关性；最后根据相似度融合规则融合各层分割结果。对比试验结果表明，该方法在有效去除杂点和噪声的同时能够较好地保留图像细节信息，并且边缘分割更加平滑，具有较高的分割精度和很好的鲁棒性。

The Analysis about SAM and Parallel Spectra Classification

Spectral angle mapper（SAM） is a physically-based spectral classification that uses an n-D angle to match pixels to reference spectra in remote sensing（RS）. In mine resource prospecting, we can use spectral vector of known mine deposit to collect spectral values that resemble ones of known mine deposit. It bases that spectra of congener things have kindred spectra. Nevertheless, there is misapplication about understanding spectral line parallelism and spectral vector parallelism. In this paper, spectral bundle has compared with SAM. RS information of mine have been extracted and synthetically analyzed and studied in factual resource prospecting. Finally, SAM effect is better than ＂spectral bundle＂.