摘要:
以1995年、2000年、2005年和2010年4期TM影像与2015年OLI影像为数据源,首先采用决策树分类算法(classification and regression tree,CART),提取山西省潞城市的土地利用信息,并进行精度评价;其次,利用土地利用变化幅度、单一土地利用动态度和土地利用程度变化综合指数等分析其土地利用的动态变化情况;然后,利用前4期数据预测2015年居民地面积,并与实际数据对比,验证GM(1,1)模型有效性;最后,利用GM(1,1)模型预测2020年的土地利用类型情况.研究结果表明,研究区土地利用动态变化情况中,林地与居民地面积增加,耕地与未利用地面积减少,水域基本保持不变;开发利用程度属于中等;土地利用结构基本保持不变;2020年研究区内耕地面积预测值为22759.32 hm2,居民地面积预测值为8854.76 hm2.
Abstract:
In this study,the authors used TM remote sensing images in 1995,2000,2005,2010 and OLI remote sensing image in 2015 as data sources,classified the image by decision tree method based on CART(classification and regression tree)to obtain the land use information of Lucheng City of Shanxi Province and did accuracy assessment. Then the dynamic change of land use was analyzed by such means as the extent of land use change, the single land use dynamics, and the integrated index of land use change degree. In addition, the GM(1, 1) model was built using first four data and was verified by the actual data in 2015. At last,the land use of Lucheng City in 2020 was predicted by using the GM (1,1) model. According to the results obtained, the forest area and the residential area increased, the agriculture area and the unused land area decreased, and the water area remained about the same in the 20 years from 1995 to 2015 in Lucheng City;the development degree achieved the medium level and the land use structure remained about the same. In 2020,the predicted value of agriculture area in Lucheng City will be 22 759.32 hm2and the predicted value of residential area will be 8 854.76 hm2.
下载: