土地利用/土地覆盖变化模型的研究

A Study on the Changing Models of Land Use/Land Cover

根据实际情况对研究区进行实验操作,对比结果为:在本研究区范围内CLUE-S模型的分类精度比BP人工神经网络分类精度高,CLUE-S模拟出来的耕地、林地、草地、水域、沙地和居民点的模拟精度百分比分别为0.032%、0.225%、53.895%、20.797%、24.683%、0.368%,总体精度可达92.08%,Kappa指数为0.9012。

The experiment is conducted according to the actual situations. Results show that in the re- search area, the classification precision of CLUE S model is higher than that of BP artificial neural net- work, the percentage of simulation precision of simulated farming land, wood, grassland, water, sand and residential areas is respectively 0. 032%, 0. 225%, 53. 895%, 20. 797%, 24. 683% and 0. 368%. The global accuracy reaches 92.08% and Kappa index is 0. 9012.