Extracting Urban Land Use Information Based On Hvperspectral Remote Sensing Data

Getting urban information quickly has great theory and practice significant at some aspects, such as developing urban planning and construction, improving urban environment, regulating urban management. The traditional method of ground investigation and measurement need too much time and human resource. An expressed and macroscopic method for updated extracting urban land use information was provided by remote sensing technology. The traditional method based on remote sensing image category extracting urban land use information was usually inaccurate. Object spectroscopy illustrated that each object has its own spectral reflectance and radiation characteristic, which can be used to distinguish different objects. Hyperspectral remote sensing can provide abundant spectral information and make up for the insufficient of spectral resolution in traditional one. It is easy to realize the precise identification and category for urban object.In this research,first,the paper illustrated the background and significance of hyperspectral remote sensing of urban surface features information extracted; introduce the basic concept of hyperspectral remote sensing; summarizes the technology of the five aspects of the hyperspectral remote sensing:illustrated the research and application status and progress of the four areas of hyperspectral remote sensing; analysis five cities typical spectroradiometer features;Finally, proposed the content and framework of research.Secondly, in this study, the hyperion hyperspectral data got preprocessing. There was dropped off bad bands,of the original242Hyperion bands,.there were149Hyperion wave bands have been retained. Then, convert the available wave bands into absolute radiation value and repair the badline. Images were atmospheric corrected using the FLAASH (Fast Line-of-Sight Atmospheric Analysis of Spetral Hypercubes) atmospheric correction software package, to convert the radiance to surface reflectance. The last step was geometrical calibration.Thirdly, the optimal bands for hyperspectral data were selected based on the specific application of urban land use information extraction. After analyzing spectral characteristic of several typical objects in whole range of wave bands, chose the range of wave bands with strong separability within study area. Find the range of wave bands with more information and less correlation after checking the information and correlation between different bands. Then, the optimal bands resampleing which contain abundant information, small correlation and strong separability for identification. The total numbers of91out of146wave bands were selected for analyzing and applying.At last, according to the specific application, the appropriate object spectral identification and classification algorithm was selected. After that, the main category of urban objects within study area would be determined. In the analysis process, the representative curve of spectrum would be selected from each object as the referenced spectrum and added into spectral database. Apply SAM software package to calculate the spectral angle of images and referenced spectrum obtained from spectral database,The image of spectral gray value can be created from it. Then setting up the threshold value in the interactive density slices tool, the different categories of objects will be identified and highlighted by specific color. At the end of the process, using regular image classification tool to have the postprocessing. The final urban object classification image was created by trying different groups of threshold value.Article by preprocessing the Hyperion hyperspectral data, the optimal band selection, and surface features identification and classification technology, analysis of the spectral separability of the nine urban surface features, and achieve better identification and classification of the nine cities in surface features. Show that the use of hyperspectral remote sensing can be more quickly and fine urban information extraction, to provide reasonable technical support and scientific basis for the implementation of urban planning and management, but also related areas of the city have some guidance and reference.