Study On Land Information Extraction Technology Based On SPOT5

The land information survey is an important basic task of the land management, and is our actual sources of accurate data about current land use. The traditional method of land use investigation not only need to expend a large amount of manpower, material resources and financial resources, but also had low efficiency. With the advance of geographic information system (GIS), remote sense(RS) and global position system (GPS), and having the high-resolution remote sensing image,it offers high efficient and reliable information and advanced technique methods.In this paper, Tailai town was taken as the study region. The SPOT5 remote sensing data was used to research on land information extraction technology. The image preprocessing from the control point, the sampling method, and the geometric correction and so on was analyzed. On the ERDAS software platform, preprocessed the SPOT5 remote sensing original data which included the geometric correction, the radiation corrections and so on.On the basis of studying images fusion technology, analyzed various fusion approaches and advanced fusion evaluating indicators. The SPOT5 panchromatic band were merged with multispectral bands through four different data fusion algorithms, which were Multiplicative transform, Brovey transform, PCA transform and Pansharp transform. We estimated the fusion images in both subjective and objective factors. Calculated the objective parameters involved mean value, variance, entropy, average gradient and correlation coefficient. Research on various unsupervised classification and supervised classification methods, then identify decision tree based on more characteristics. Analyzed the bands to determine bands combination (spectral characteristics); computed the normalized difference vegetation index (NDVI); Merged images which adopted pansharp transform were used to image classified automatically. On the ARCGIS software platform built the DEM model based on the paper map to derive terrain features. Analyzed texture features on ENVI software platform. Choosing sample data, constructed decision tree on see5.0 platform, and then extracted the classification rules. At last, classification precision was determined according to evaluation indexes.It can be shown from this research that the fused images had higher spatial resolution while maintaining the rich spectral information. And the visual effect and the accuracy of the classification had been greatly enhanced. It proved the Pansharp transform applied to SPOT5. In view of SPOT5 with high spatial resolution, the general classification is no longer applicable, and precision of the classification can't meet the requirements. So introduced the idea about decision tree and considered multi-feature to build the decision tree. The multi-feature included spectrum, vegetation indices, the terrain feature and the texture features. The image was classified according to the decision rules. It can be seen that the classification accuracy has greatly improved.