Application research of MODIS data in monitoring land use change in Fujian

摘要

It was necessary and significant to explore the low-cost, high-precision and real-time access method of land-use/cover using the MODIS data multi-temporal and multi-spectral for quickly assess regional land use/cover change. Firstly, the study used maximum values compose (MVC) to select the optimal MODIS data in Fujian study area because there were mountainous landform and cloudy climatic condition in study area. Secondly, the characteristic variables of surface albedo, vegetation index(NDVI), water index(NDWI) and so on were combined, Moreover, the decision tree classifier system was established based on multi-factors composition for land cover/use classification in Fujian province. The results showed that the decision tree classifier was better than conventional maximum likelihood classifier, and was well applied the MODIS data to classify the land-use/cover of Fujian, because the decision tree classifier system took advantage of the MODIS multi-spectrum characters and artificial intelligence and could achieve a certain high precision, which made an important impact on monitoring land change and protecting arable land.