基于PLS-PP的延吉市土地利用变化趋势预测研究

利用延吉市1992年和2007年两期Landsat TM数据,在分析土地利用变化的基础上,应用偏最小二乘回归的投影寻踪耦合（PLS-PP）模型,对延吉市的土地利用变化进行了分析和定量预测。研究结果表明,偏最小二乘回归的投影寻踪耦合（PLS-PP）模型与传统的偏最小二乘回归（PLS）模型方法比较,预测值精度显著提高。根据PLS-PP模型的运算结果,按照目前的变化速度,延吉市林地、草地将逐渐减少,耕地、居民点工矿用地、未利用地、水域用地逐渐增加,其中未利用地、水域用地变化幅度较大。该方法可为土地利用变化预测研究提供一种新的思路,预测结果可为土地管理部门制定相应的对策提供依据,从而达到土地资源优化配置与合理利用的目的。

Analysis on Prediction of Land Use Change Based on PLS-PP in Yanji City

Using the remote sensing data of TM satellite image data between years of 1992 and 2007,On the basis of the analysis of land use change,PLS-PP model was built to analyze and predict land use change in yanji city.the computation results showed that the relative error of PLS-PP model was smaller than traditional PLS model′s,and it had improved the prediction precision evidently.According to the result of PLS-PP,on condition that the rate of land use change was not changed,area of woodlands、grasslands would decrease,but area of cultivated land、residential and industrial or mineral land、unused land and water area will increase.the water area and unused land changed significantly.This research could provide a new way to predict land use change,and the result could provide a basis for land management department to formulate land-use policies,then attained objective to optimize and rational use the land resources.