Scale Effect Analysis Of Regional Vegetation Coverage Extraction

Vegetation coverage is the important factor which represents the growth condition or ecological environment or soil erosion.Increasing of vegetation coverage calculation precision in the area can provide comprehensive control of soil and water loss.Remote sensing estimation method is the main technical means which is used to calculate vegetation coverage,moreover,it alos offers variety of multi-scale databases and brings scaling effect at the same time. What's more,it has great significance that make the discussion about the scaling effect generated by various resolution for improving the precision of vegetation coverage calculation. Different space within the scope of vegetation coverage information measurement has the suitable scale of remote sensing, The loess plateau is severe soil and water loss area in our country, researching on different regional choice of appropriate scale remote sensing data to estimate vegetation coverage not only can improve the precision of vegetation coverage, but alos save the cost, and provide scientific comprehensive harness of soil and water conservation in loess plateau with referable scaling and quantitative data.The TanHe Ruver Basin as the research area in the loess plateau, selecting multiresolution remote sensing images, including high resolution image data such as Worldview (1.8 m), IKONOS (4 m), RapidEye (5 m), ALOS (10 m), moderate resolution imaging is the Landsat8 (30 m) and low resolution MODIS images (250 m) and Spot Vegetation (1000 m). On the basis of multi-source remote sensing data for the extraction of the optimal Vegetation Index, analyzing the scale effect of Vegetation Coverage inversion and exploring the variation of the vegetation coverage through the descending on the resolution. To make exploration of entire area in Yanhe basin to research the suitable resolution of remote sensing image for extracting Vegetation Coverage. The preliminary results obtained as follows,First Point, the analysis of the vegetation index selectionBased on imaging time of remote sensing with the measured season, selected the five kinds of commonly used green vegetation index, respectively is the normalized difference vegetation index (NDVI), atmospheric impedance difference vegetation index (ARVI), Modify Soil Adjusted Vegetation Index (MSAVI), enhanced vegetation index (EVI) and perpendicular vegetation index (PVI). Analysis of multi-spectral remote sensing data extraction of vegetation index of regional applicability, with actual coverage through correlation analysis, the result concluded that he optimal vegetation index of the inversion yan river basin in the vegetation coverage is atmospheric impedance difference vegetation index (ARVI).Second Point The scale effect analysis of the regional vegetation coverage inversion selection of vegetation distribution more uniform 50 sq km area based on YanHe River Basin,and had calculated the vegetation coverage from the multisource remote sensing image.There existed obviouse difference vegetation coverage result between different image resolution.Proceeding calculation of vegetation coverage from variety resolution images based on 50sq km homogeous vegetation area. The minimum value increased, the maximum to reduce gradually, weakening the effect of extraction of vegetation information.30 mlandsat8 ARVI and average coverage are lower than the high resolution image data, MODIS (250 m) and Spot Vegetation (1000 m) images to extract the region of Vegetation information ability is bad, there is lack of Vegetation coverage information, does not apply to information extraction of Vegetation in this area.Third Point Regional vegetation coverage inversion accuracy analysisIn the Worldview (1.8 m) for reference to data analysis for coverage extraction of vegetation species in 50 square kilometers RapidEye (5 m), ALOS (10 m) and Landsat8 (30 m) three dimension data coverage results accuracy. Using transfer matrix of land use/landscape is commonly used in geological, respectively establish a multi-scale vegetation coverage of the video data transfer matrix, analysis of different resolution image data coverage as follows, it is concluded that the scale of remote sensing data coverage accords with the true value of the ratio of:RapidEye, ALOS,37.17% and 42.39%,34.69% 34.69%, MODIS Landsat8; Which can be obtained within 50 square kilometers of 10 m resolution ALOS data inversion of vegetation coverage of the highest precision.Fourth Point remote sensing data extraction scale research of region vegetation coverage information of areas suitable for inversion of the vegetation coverage of multi-scale resolution image data to send a sample, using the method of local variance of statistical resampling multi-source data after the resolution of the vegetation coverage information, reach 50 square kilometers, within the scope of the resolution than 30 m rough (more than 30 m) space scale of remote sensing data extraction of vegetation information will appear in the process of a large number of missing, unable to ensure accuracy.This law, carry out the whole yan river basin the coverage to extract the resolution suitability, based on data from the full yan river basin (including Landsat8, MODIS, Spot Vegetation), in accordance with the requirements of the accuracy of the premise, can inversion yan river basin in the Vegetation coverage is appropriate minimum resolution can be reduced to 300 m-400 m.