传统的与光谱反射率匹配的矿物填图方法受图像质量、 大气及环境背景等因素影响较为明显, 矿物识别效果较差。 针对上述问题提出一种基于多类型光谱特征参数匹配的矿物信息识别方法, 综合利用多种光谱特征参数在大气及环境背景变化中的相对稳定性特点, 实现矿物信息的高精度识别。 选取美国内华达州Cuprite矿区AVIRIS机载可见-红外光谱成像仪高光谱数据, 以USGS美国地质调查局波谱库中的矿物波数为参照光谱, 分别计算AVIRIS影像和USGS波谱库中典型矿物的光谱特征参数, 在对多种光谱特征参数综合分析的基础上, 构建Cuprite矿区矿物识别模型, 进行矿物填图实验。 最后使用Clark等在该地区的矿物填图结果数据, 对实验结果进行了验证, 结果表明: 该方法矿物识别结果与Clark等的填图结果吻合度高, 整体矿物识别精度达到78.96%。

The traditional mineral mapping methods with remote sensing data, based on spectral reflectance matching techniques, shows low accuracy, for obviously being affected by the image quality, atmospheric and other factors. A new mineral mapping method based on multiple types of spectral characteristic parameters is presented in this paper. Various spectral characteristic parameters are used together to enhanced the stability in the situation of atmosphere and environment background affecting. AVIRIS (Airborne Visible Infrared Imaging Spectrometer) data of Nevada Cuprite are selected to determine the mineral types with this method. Typical mineral spectral data are also obtained from USGS (United States Geological Survey) spectral library to calculate the spectral characteristic parameters. A mineral identification model based on multiple spectral characteristic parameters is built by analyzing the various characteristic parameters, and is applied in the mineral mapping experiment in Cuprite area. The mineral mapping result produced by Clark et al. in 1995 is used to evaluate the effect of this method, results show, that mineral mapping results with this method can obtain a high precision, the overall mineral identification accuracy is 78.96%.