The Research Of Material Information Extraction Methods And Applications On The Lunar Surface Based On Multisource Remote Sensing Data

Detection of lunar surface material information is an important part of the lunar exploration, the inversion of lunar surface material information could provide important parameters for the study of the origin of the moon and evolution history, which is very important significance for lunar exploration, man’s landing on the moon and the exploitation of moon’s resource. The research based on lunar samples does not represent the whole moon because of limited sample quantity and distribution area. Remote sensing is ability to quickly get the data of multispectral and hyper spectral. Especially the hyperspectral remote sensing data has fine spectral information. The extraction of material composition information on the lunar surface based on the spectral dimension, that plays an important part in detection of lunar surface material information. With the vigorous development of lunar exploration, we have access to a lot of multispectral and hyperspectral data, but the quantitative analysis of spectral signatures on the lunar surface still needs further development. Therefore, the research of the spectral features of material information on the lunar surface and spectral feature information extraction method is fundamental problem of hyperspectral detection. Analyzing the characteristics and difficulties of detecting lunar surface material and building method system of the lunar surface material detection based on Multisource Remote Sensing data, that is the important prop and technical support for further application of lunar satellites data.The paper put the application of lunar remote sensing data as entry points, extract synthetically the lunar surface material information based on multisource remote sensing data, carry out the spectrum characteristic features and extraction methods and application research of lunar material information. Research results provide a reference for the applications of lunar remote sensing data and provide technical support for the exploitation of moon’s resource. Main contents of this paper are as follows:1. Analyzed mainly spectrum characteristics of the lunar rocks and minerals in the VIS-NIR and the influence factors of the lunar material spectral response.Using the lunar sample spectra data of RELAB spectral library, spectrum characteristics of main rock and mineral are analyzed. Based on the analysis, the paper researches the influence factors of spectral response on the lunar surface and comprehensive evaluation index system of spectral response factors.2. Because the extractions of remote sensing information need accurate reflectance, the paper has studied on preprocessing of remote sensing data and the inversion methods of reflectance.The paper carried on the thorough research on preprocessing of selected remote sensing data, including bad point detection and repair, bad line detection and repairs and remove stripe. The lunar surface reflectance inversion methods are studied and choose the method suitable for remote sensing data of the moon.3. The paper studied on the methods of spectral feature selection and extraction in depth, and is mainly focused on the spectral feature extraction method based on wavelet decomposition and studied the method for determining the best wavelet decomposition scale.The paper study the existing technologies of spectrum feature selection and extraction, use these methods for lunar sample spectra data, is mainly focused on the feature extraction method based on wavelet decomposition and determine the best wavelet decomposition scale on the spectral feature extraction.4. Extraction of Lunar material information methods based on multispectral and hyperspectral was studied. Experiments of lunar information extraction have been carried out.The paper research on methods of the lunar surface information extraction, including lunar material inversion based on band features, information extraction based on spectral feature matching, spectral decomposition technique of the mixed pixels and information extraction based on target detection algorithms. Extraction and identification of the lunar surface material information is studied based mainly on CE-1 ⅡM and Chandrayaan-1 M3, auxiliary on Clementine UVVIS-NIR and SELENE MI.5. Put up digital lunar information sharing service platform.The paper put up the lunar information comprehensive sharing service platform based on G/S model, we can use the platforms to effectively analyze on material information by multiple source information and improve the utilization efficiency of the remote sensing data.Based on the research contents in this paper, the main innovation points are as follows:1. In this paper, a method which is based on multi-source remote sensing data to extract comprehensive lunar material composition information is proposed.The paper researches spectral mechanism of lunar material and analyzes the spectrum characteristics of main lunar rock and mineral based on laboratory spectral data of lunar return sample, then analyzes and summarizes data preprocessing and the application research of lunar information extraction based on mainly hyperspectral data (CE-1 IM and Chandrayaan-1 M3) and auxiliary multispectral data (Clementine UVVIS-NIR and SELENE MI). The key technology of application and the validity of the data are analyzed. The paper takes Apollo 16 sampling site and Aristarchus impact crater as examples and makes the retrieving of TiO2, FeO and maturity content, and verify the effectiveness of the inversion model.2. The paper proposes an evaluation index system of lunar material spectral response based on hierarchy model.The paper considers various factors of the lunar material spectral response. Comprehensive evaluation index system of spectral response factors is tried to design. It provides scientific basis to for realizing quantitative extraction of lunar material information. Evaluation index system can evaluate the influence of various factors. It makes extraction of the lunar material information more accuracy.3. The paper proves that the wavelet decomposition is available for spectral feature and put forward a method by computing correlation coefficient abrupt change between approximation signals and the original signal to determine the optimal decomposition scale of wavelet.The paper uses wavelet decomposition and reconstructs to experiment the spectral feature extraction of lunar samples. The result proves the feasibility of wavelet decomposition in the spectral feature extraction. It is the emphasis and difficulty in using wavelet analysis spectral feature how to select the best wavelet decomposition scale. The paper uses the correlation coefficient abrupt change of the approximation of the decomposed signal and original signal as the choice of optimal scale, and the experiment proved that the method has a good effect in spectrum feature extraction of lunar mineral.