利用目标向量空间变换方法提取含矿岩层的研究——以新疆卡特里西地区为例

STUDY OF MINERAL-BEARING ROCK FEATURE EXTRACTION USING THE TARGET-SCALAR TRANSFORMATION ALGORITHM ——AN EXAMPLE FROM KATELIXI, XINJIANG

多因子线性变换是处理多波段遥感数据的数学方法 ,应用该方法对于图象主要特征的分离和提取具有很好的效果。但是 ,遥感图象数据中的岩性信息光谱分布复杂多样 ,在信息强度和分布方面较弱 ,使用现有的变换很难将岩性信息提取出来。为了完成遥感岩性填图的任务 ,本文提出利用目标向量空间线性变换的方法 ,简化所要提取的目标信息的光谱结构 ,提取目标信息。根据研究区内岩性目标地物的图象灰度光谱分布选择空间域坐标轴的位置 ,使需要分离的岩性信息只在一个轴上的投影不为 0 ,从而将岩性目标信息分离出来。该方法在新疆且末卡特里西地区提取铜矿带岩性取得了较好的结果。

The algebra transform method is very effective to process the multiple band TM data, such as the Principle Analysis in feature space and the Tasseled cap transform for vegetation identification. The rock information is very complex among the information composed in the TM data. In order to mapping the rock area, a new target scalar transformation is tested for the weak rock information extraction. The method is setting the coordinate of the scalar space at the selected target. The transformation acts on the original data, so that the interested information is exposed in the composite. The method achieved a good result in Katelixi copper mine area, XinJiang Uygur Autonomous Region.