融合主成分分析与光谱角匹配的金铜矿遥感探测方法

Gold and Copper Deposit Remote Sensing Detection Method Based on Principal Component Analysis Method and Spectral Angle Matching Method

针对金铜矿高光谱遥感探测中存在的数据量大、数据处理繁琐等问题,融合主成分分析与光谱角匹配方法,提出了一种金铜矿遥感探测方法。该方法首先采用主成分分析法对金铜矿高光谱遥感数据进行降维处理,然后利用光谱角匹配法对目标光谱进行匹配,从而实现对金铜矿进行遥感探测的目的。在福建紫金山金铜矿区进行实地取样,进行可见光—近红外光谱测试分析,采用所提方法对高光谱遥感数据进行处理,并对样品进行了金、铜矿及围岩识别试验。结果表明:所提方法能够有效简化金铜矿遥感高光谱数据处理流程,使得实测遥感光谱原始数据从973个波段简化为6个主成分,利用该6个主成分进行光谱角匹配时,大大简化了计算过程,且光谱信息损失较少,岩(矿)样品识别成功率达到了74%,表明该方法在金铜矿高光谱遥感探测方面有一定的适用性。

In view of the existed complicated problems of the large amount of remote sensing data and complex data processing flow in the process of hyper-spectral remote sensing detection of gold and copper deposit,in order to improve the efficiency of the remote sensing detection of gold and copper deposit,a new remote sensing detection method based on principal component analysis method and spectral angle matching method of gold and copper deposit is proposed. Dimension reduction of the hyper-spectral remote sensing data of gold and copper deposit is conducted by adopting the principal component analysis method; then,the target spectrum is matched by using the spectral angle matching method to realize the goal of the remote sensing detection of gold and copper deposit with high efficiency. Taking the Zijinshan gold and copper deposit as the study example,the samples of gold,copper deposits and surrounding rocks are collected,the experiments of testing the samples with visible-near infrared spectral by using the detection method proposed in this paper is done to dispose the hyper-spectral remote sensing data and identify the gold,copper deposit and surrounding rocks. The experimental results show that the processing flow hyper-spectral remote sensing data of the gold and copper deposit is simplified effectively by using the detection method proposed in this paper,the measured spectral remote sensing data is simplified into 6 main components from 973 bands which can facilitate the flow of computing with little spectral information loss in spectral angle matching,the success rate of the detection method proposed in this paper of the gold and copper deposit is overall reached to 74%,it is further indicated that the detection method has a better applicability in gold and copper deposit in the prospecting work by using hyper-spectral remote sensing technology.