基于SAM与SVM的高光谱遥感蚀变信息提取

高光谱遥感技术的发展，提高了遥感技术的定量化水平，要求人们从光谱维去理解地物在空间维的变换。提出了一种光谱角匹配技术（Spectral Angle Mapper，SAM）与支持向量机（Support Vector Machine，SVM）相结合的高光谱遥感蚀变信息提取模型，在光谱维提取地表的蚀变信息。鉴于SAM算法仅考虑波谱矢量方向，忽略辐射亮度大小的缺点，利用SVM算法对SAM的提取结果进行二次分类，利用网格搜索法并结合分类精度评估进行参数寻优。通过AVIRIS高光谱数据实验证明，提取的蚀变信息分类精度为78.172 6%，Kappa系数为0.712 5。该模型计算方便，对于解决光谱维的地物分类及相似矿物的蚀变信息提取具有一定的实际意义。

Extraction of hyper-spectral remote sensing alteration information based on SAM and SVM

With the development of hyper-spectral remote sensing technology, the level of quantitative remote sensing technology has improved. Aiming at the hyper-spectral image cube, the understanding and data processing in image spatial dimension must be changed to that completed in the spectral dimension. Therefore, an image classification model combined with SAM（Spectral Angle Mapper） and SVM（Support Vector Machine） is introduced, and extracts alteration information in the spectral dimension. In view of the SAM algorithm considering only the spectrum direction, ignoring radiance size, the second classification is made for the SAM results using SVM algorithm and the best parameter is sought using grid search method combined with the classification accuracy assessment. The results of AVIRIS hyper-spectral data show that the classification precision of alteration information reaches 78.172 6%, and a Kappa coefficient of 0.712 5. This model is convenient calculation, and has some practical meaning in solving spectral dimension terrain classification and similar mineral alteration information extraction.