Research On The Analysis Of The Spectrum Data And Extraction Metheds Of Minerals

Hyperspectral remote sensing is one of important high-tech and application technologies in the remote sensing area, especially in Geo-remote sensing area from home and broad. Due to the limitation of the spectral resolution in traditional Multi-spectral remote sensing, it is complicated to recognize the special but rich spectral information of the ground objects. From the 80s' of last century, hyperspectral remote sensing is wide used in mine search, the extraction of the spectral information and so on, also we get so many successful results. The research shows that, it is possible to extract the indicative spectral information of the geological body with the assistant of the Nano-hyper remote sensing dataCurrently, the Geo-application and research of the hyperspectral remote sensing in home and broad is mainly collected at several parts as follows: first, the analysis of the spectral character of minerals. Most of the research on the characteristics of the minerals is mainly concentrated on visible light and near-infrareds whose wave length are 0.5-2.5μm, whereas the research shows that a lot of minerals own less obvious wave characteristics, nevertheless, they are well- shown under the mid-far infrareds (such as Almandite). Therefore, it is necessary to systematically analyze and research the mid-far infrared wave characteristics of the minerals. Second, test analysis of the wave characteristics of the minerals. Traditional wave test is limited by hardware and software of the instrument, also limited by the communication condition. High precision wave test was only applied in a few departments. Currently, with the instruction of the portable high precision Spectrometer, the remote sensing worker is quite familiar with wave test so that the mineral spectrum work is widely promoted. Third , the research on the extraction algorithm of the mineral spectral wave information and inversion model. Although there are many the extraction algorithms of the mineral spectral wave information and Inversion model, but to the extract the weak information of the geological body, there are some faults in the existent methods, it can not resolve some application problem, so it is necessary to create the new algorithm and model.Throughout the application of geo-remote sensing research from home and abroad, hyper remote sensing still stays at the test level. Although we get some result, it is still away from the application level. So to some degree, the forming mechanism of the ground object hyperspectral remote sensing, the spectral test and the influence, the methods of hyperspectral data and the technology of the remote sensing images of hyper spectrum processing and information extraction and application is not very developed. Based on comprehensively analyze the disadvantages of above research and specifically analysis the mid-far infrared spectral characters, the author designs the new inner test method of the mineral spectrum. Moreover, the author promotes the existed algorithms and models, and offers the new dimensionality reduction and matching method. At last, through the application of the examples, the author offers the mineral information of the technology process of the hyper remote sensing data.The main research and creative consequence of this article as follows:(1) This article mainly analyzes spectral band caused by the vibration process of the molecular and crystal. Based on the group theory analysis, through the matrix, the author projects the number of the mineral spectral absorption band caused by the basic frequency vibration. Above this, the author analyzes the spectral characteristics of different kinds of minerals. With the thinking of the few analysis and statistics of the spectral characteristics of the minerals the mid-far infrared wave in out country, this article shows the characteristics of mid-far infrared band of the minerals, also it widely expend the existed research range of the characteristics of the mineral spectral. It owns the great reality sense to the set and selection of the RS sensor.(2) Spectrum test always is foundation of the rock-mineral identification and the ground object recognition in remote sensing. With the development of the portable spectrometer and the technology of the hyper remote sensing, the mineral spectral test becomes more and more important. The author of this article designs inner mineral spectral test method. With the application of the portable ASD spectrometer, the author finished the spectral test of some typical minerals in the experimental area. Also, based on the C# environment of Visual Studio 2005 and Microsoft Access database technology, initially establishes spectral data management system. It is new way of thinking of access and management of the inner mineral spectrum.(3) The author offers the method of the mixture dimensionality reduction. Different from the exeistent methods, during the sub space decomposition of the ground object, the method of the mixture dimensionality reduction take the unique spectral characteristics,especially the Diagnostic spectrum characteristics as an independent sub space, this can reduce the loss of information of the unique spectral characteristic of the ground object.Through the analysis of the experiment, this article depicts this method is more advanced than the method depended on the band characteristics to the sub space division.(4) The author offers the evaluate method about the similarity which based on the spectral ratio and margin.Based on the method, the author obtains the statistics of the ratio and margin of the spectrum. Depending on some relative indicators(the max and min deviation, also the average deviation), the author judges the matching rate of the spectrum. This method enrich the existed the technology of the spectral matching. It also owns the prospects of the application and the value of the promotion.(5) Through the application research on the extraction of the mineral information RS data, it initially forms the practical technology process of the mineral information recognition and extraction of the hyper and multiple spectrum RS image based on the spectral data.