Change detection of land use and land cover of the southeast coastal area in Fujian, China

Land use and land cover (LULC) can usually indicate the change of the eco-environment for one area. In this study, a knowledge based decision tree classification system based on the spectrums knowledge of ground objects and their spatial knowledge was developed for the research area, the southeast coastal area in Fujian province, China. The knowledge based classification system was built by using Spatial Modeling Language and ERDAS Macro Language provided by ERDAS IMAGINE software as secondary development languages. Two scenes of Landsat TM imagery acquired in 1986 and 2000, respectively, were classified into 12 classes by use of the system. Then the LULC change was detected by comparing the classification results from two different imagery data. The results reveal that large-scale changes of the LULC types represent the eco-environment degradation. By field checks, the classification output matched well with the reality (natural conditions) and can guide the further work within the research area.