论文部分内容阅读
光谱角度填图（ＳＡＭ）是成像光谱图像处理技术中，对岩石矿物进行谱形识别的主要方法之一，但其应用效果并不十分理想。以内蒙古渣尔泰山为实验区，从岩石实验室反射光谱的相似系数聚类分析入手，研究了光谱角度填图方法的优劣，及其产生的原因。探讨了利用均值图像参与分类的处理方案，从而改善了光谱角度填图方法中易于产生“异类同谱”的不足，获得良好的岩性识别效果。
Spectral angle mapping (SAM) is one of the main methods for spectral identification of rock minerals in imaging spectral image processing technology, but its application effect is not very satisfactory. Starting with the clustering analysis of similarity coefficient of reflection spectrum in rock laboratory, the experiment was carried out in Zherl Mountain, Inner Mongolia, to study the advantages and disadvantages of spectral angle mapping method and its causes. This paper discusses the scheme of using average image to participate in classification, which can improve the defect of “heterosexual spectrum” easily generated in spectral angle mapping method and obtain good lithology recognition effect.