(1. 中南大学 有色金属成矿预测教育部重点实验室，长沙 410083；
2. 中南大学 地球科学与信息物理学院，长沙 410083；3. 青海省地质调查院，西宁 810012；
4. 湖南有色金属控股集团有限公司 矿山管理部，长沙 410015；5. 河南大学 环境与规划学院，开封 475001)
摘 要: 褐铁矿所形成的铁帽是寻找铝土矿的重要标志。在分析铝土矿地质特征、光谱特征和遥感影像特征的基础上，引入独立分量分析(ICA)法用于铝土矿的蚀变信息遥感提取，并与传统的主成分分析(PCA)法和比值法进行了比较。结果表明：3种方法的准确率分别为84.5%、80.9%和79.4%，并且独立分量分析的第一分量IC1的值越低，铝土矿的蚀变程度越高，信息提取的准确率也越高。遥感提取的铝土矿蚀变面积约210.1 km2，可划分为3个区带，主要分布于海拔250 m以上且切割相对较大的山顶和斜坡上，而低海拔的平缓地带铝土矿基本不发育。初步查明了区内铝土矿的地表分布规律，为进一步的找矿勘查工作提供有效的指示信息。
关键字: 铝土矿；蚀变；独立分量；信息提取；成矿预测
remote sensing data based on independent component analysis
（1. Key Laboratory of Metallogenic Prediction of Nonferrous Metals, Ministry of Education,
Central South University, Changsha 410083, China;
2. School of Geosciences and Info-Physics, Central South University, Changsha 410083, China;
3. Qinghai Institute of Geological Survey, Xining 810012, China;
4. Department of Mineral Resources, Hunan Nonferrous Metals Holding Group Co., Ltd., Changsha 410015, China;
5. College of Environment and Planning, Henan University, Kaifeng 475001, China）
Abstract:Iron cap formed from limonite is an important prospecting symbol of bauxite in the studied area. Based on the feature analyses of geological, spectral and remote sensing images, independent component analysis (ICA) was used to extract the alteration information of bauxite. The traditional methods of the principal component analysis (PCA) and ratio were compared with ICA. The test results show that the extraction accuracies of ICA, PCA and ratio are 84.5%, 80.9% and 79.4%, respectively. The lower the value of the first component of ICA is, the higher the degree of bauxite alteration and the extraction accuracy are. The surface area of bauxite alteration extracted from IC1 is about 210.1 km2, and is divided into three zones. It mainly distributes in the hill and slope where the elevation is higher than 250 m. There is little bauxite in the flat zone with low elevation. The distribution of bauxite in this area is verified preliminarily, which provides very effective instruction information for further exploration.
Key words: bauxite; alteration; independent component; information extraction; metallogenic prediction
