Font Size: a A A

The Study On The Landscape Pattern Of Karst Area Landuse Based On CBERSO2B And The SAR Data

Posted on:2010-01-04Degree:MasterType:Thesis
Country:ChinaCandidate:H M LvFull Text:PDF
GTID:2178360275982564Subject:Physical geography
Abstract/Summary:PDF Full Text Request
Most parts of Guizhou is the model of the karst environment, the environmental capacity is small and the ecological environment is frangible,In order to study the ecological environment better in the karst area,this article took Bijie city in Bijie area as the typical type area,selecting the CBERS02B and the SAR datas for the first time,discussing the imagery processing technical usability about the two types of remote sensing images in the karst mountainous area as well as the landuse ideal Classifing approaches.On the processing technology of the remote sensing images, this research had used methods of Speckle Suppression,the Edge Enhancement,the Wallis Adaptive Filtering,the Histogram Equalization and the Haze Reduction to the SAR and CBERS02B datas.During the processing to the SAR,it discovered that the Mean filter and the Lee filter reduced the speckle noise better; the Robinson the 3-level model and Prewitt the Gradient model strengthened the edge effect of the images greatly;the Wallis Adaptive Filter enhanced the images in obviously different effect under the different window sizes,when the window was bigger,the luminance information was more centralized.AND during the processing to the CBERS02B,it also discovered that the Histogram Equalization strengthened the brightness of the whole images very well,and after the low frequency ambiguity and high frequency ambiguity processing,the transportation land texture information had been strengthened.The method of the PCI histogram matching had demonstrated the good effect in resoluting merge of the SAR and in the CBERS02B and had also removed the original multi-spectrum cloudy layer information very well.This article performed supervised classification to the CBERS02B,and basing on the supervised classification, plusing the texture classification of the radar single wave band,comparing the precision of the classifications,studied the landscape pattern about the landuse in the area and under the different rocks,finally it drew conclusions as follows:(1),The pure spectrum classified precision of the CBERS02B had been improved to a certain extent after plusing the texture classification of the radar.(2),The fractal dimension value of the water was the biggest,and the bare rock gravel's was the smallest,which had demonstrated the human activity acted greatly to the bare rock gravel;the entire research region had also demonstrated that the relative rich degree and broken degree were lower, evenest was smaller,complex was bigger as well as higher superiority,with the farmland,the forest land occupying the dominant position.(3),The forest and the farmland had occupied the dominant position whatever in the karst,half-karst and the non-karst region,and the garden landuse was much more broken than others; The lawn,the garden landuse and the waters was much more broken in the karst;The farmland and the construction land of the artificial landscape were much more broken in the half-karst, and the integrity of the natural landscape's forest was good in the non-karst;The bare rock gravel was too much bigger than other two kinds in the karst area percentage,the non-karst region had demonstrated that the relative rich degree and broken degree were lower,it fell under the one kind or the minority several kind of landscape control's phenomenon was much more obvious than the karst and half-karst;The half-karst region had demonstrated that the evenest was bigger,complex was smaller as well as lower superiority,with higher broken degree...
Keywords/Search Tags:CBERS02B, SAR, karst, landscape pattern
PDF Full Text Request
Related items