Synthetic analysis of urban landscape pattern and environmental impact based on Remote Sensing - a case study of Xuzhou City

摘要

Taking Xuzhou City as a case, the studies of landscape classification, pattern analysis and eco-environmental impacts evaluation are made using Landsat TM/ETM+ images in this paper. For this purpose, firstly urban landscape information is extracted from Landsat TM/ETM+ images using C5.0 decision tree classification method based on spectral features, texture features and shape information. The results prove that the classification accuracy can meet the requirements of landscape patterns analyzing and evaluating. Then according to classified images, urban landscape pattern is analyzed based transfer matrix and landscape pattern metrics at the scale of landscape and class. Finally, based on the analysis, urban landscape pattern impacts on environment are evaluated qualitatively from three angles: heterogeneity, fragmentation, and edge effect. The results, which indicate that the level of fragmentation, diversity and connectivity of Xuzhou landscape in 2005 is higher than in 1987, but the level of congregation in 2005 is lower than in 1987, can reflect the urban expansion and the change of landscape patterns effectively.