Based On Modis Data, Information Extraction And Land Cover Change Research

Using remote sensing data to do the extraction of the land use and cover is a hot topic in global change research, and MODIS data has been widely used by its ease of access to information and timeliness. This paper choose 2002-2008 MODIS 09 A1 data as research data and west of Sichuan Province as experimental area to do the method of extraction innovations and the main land cover changes research.The core idea of combination extraction based on multi-temporal NDVI is extraction the vegetation by their time change process. The vegetation phonological theory is: Because the same vegetation has different growing when they are in different seasons or different periods, their performance in remote sensing images will appear a certain rule changes, especially the NDVI value. Different vegetation has different NDVI value. The realization is: reclassification and reorganization the multi-temporal NDVI to get the sequence and its corresponding number of pixels; Through the vegetation sampling statistical analysis to determine the types of vegetation, so that can achieve the extraction of vegetation information. This method carried out in West of Sichuan province and the results show that: the total accuracy is 72%, area accuracy is 85%.Compared with supervise classification, using this method is better. To Achieve automatic classification by programming.Green vegetation and the snow is the main land cover in western Sichuan. This paper based on the 2002-2008 NDVI sampling feature points and reference to the international standard to determined the green vegetation extraction model. Using this model to extract Green vegetation throughout the year 2002 and synthesizing green vegetation time covered map by 24-day intervals. The results show that: the green vegetation cover in Liangshan and Panzhihua area is better than Ganzi region; especially more than 60% of Panzhihua area the green vegetation cover more than 285days, and then analyzed the green vegetation covered in western Sichuan; Utilizing MVC method to synthesis largest area of green vegetation of 2002-2008 and using the extraction model to extract the green vegetation to find out: the changes of green vegetation is bigger in seasons and smaller in years. The 7-year green vegetation's area analysis shows an upward trend in basic. Based on Spectroscopy to build the snow extraction model .Using this model to extract snow throughout the year 2002 and synthesizing snow time covered map by 24-day intervals. The results showed that: the western Sichuan's snow changes a lot in different seasons and 90% of the snow cover time is less than 100 days. To calculate the max snow cover area of the 7 years. Comparative analysis results indicate that: In recent years the area of snow shows smooth fluctuations at 9000 Square kilometers and the max area of snow cover is 11000 Square kilometers. The snow of western Sichuan is significant affected by the Qinghai-Tibet Plateau climate.