土地利用变化遥感监测精度评价系统——以随机抽样为基础的方法

The accuracy assessment system of remote sensing monitoring land use change: a method based on random sampling

中国正在采用遥感技术对 5 0万以上市区非农人口城市的建设用地扩展情况进行分期监测 ,监测结果的精度评价是该项工作的一个重要方面。本文提出以分层随机抽样的方法来获取对遥感监测的精度分析 ,内容包括监测精度评价的指标系统、分层随机抽样样本容量的确定和在各层的分配、抽样评价的工作步骤以及抽样调查的统计分析和结果表达 ,并以北京和重庆 2 0 0 1年的遥感监测数据和实地调查数据为基础进行了精度评价的实际演练。

The shrinkage of China’s cultivated land is a world-wide concern, which is believed to arise from the increase of land used for construction purpose. In order to obtain a timely understanding of the land expansion for urban construction and the decrease of cultivated l and in China, remote sensing is being employed to monitor the expansion situation of land used for construction in the cities having above 500,000 non-agricultural population during various periods, and the accuracy estimation of the monitoring results constitutes an improtant aspect in this task. Up to now, many estimation indexes have been proposed which serve as a favorable base for future research. However, there still exist some drawbacks to be solved.In the article, a method of stratified random sampling is presented to obtain the accuracy analysis of remote sensing monitoring.The accuracy assessment is comprised of technological and regional assessment. It involves the assessment system of accuracy monitoring, the capacity determination of the stratified random sampling method, the procedure for sampling and evaluation, and the statistic analysis and result expression of the random survey. In addition, to verify the feasibility and suitability of this kind of method,the practice of the above system, based on data for Beijing and Chongqing in 2001 collected by remote sensing and fieldwork, is performed.Our work successfully applied the sampling theory to geoscienlific survey, which is of great theoretical and practical significance.