Optimal decision-making model of spatial sampling for survey of China's land with remotely sensed data

查看全文 作　　者：LI Lianfa1,2, WANG Jinfeng1 & LIU Jiyuan1 1. Institute of Geographic Sciences & Natural Resources Research, Chinese Academy of Sciences, Beijing 100101, China;2. Department of Intelligence Science, Peking University, Beijing 100087, China Correspondence should be addressed to Li Lianfa 出　　处：《Science China Earth Sciences》索引2005年第48卷第6期,共13页 基　　金：the National Key Fundamental Research Development Planning Project(Grant No.KZCX1-Y-02);the High-tech Research and Development(863)Programme of the Ministry of Science and Technology(Grant No.2002AA135230);the Projects of the Chinese Academy of Sciences(Grant Nos.KZ951-A1-302,KZ951-A1-203, KJ951-B1-703); the National Natural Science Foundation of China(Grant Nos.49871064 , 69896250). 摘　　要：In the remote sensing survey of the country land, cost and accuracy are a pair of conflicts, for which spatial sampling is a preferable solution with the aim of an optimal balance between economic input and accuracy of results, or in other words, acquirement of higher ac-curacy at less cost. Counter to drawbacks of previous application models, e.g. lack of compre-hensive and quantitative-comparison, the optimal decision-making model of spatial sampling is proposed. This model first acquires the possible accuracy-cost diagrams of multiple schemes through initial spatial exploration, then regresses them and standardizes them into a unified ref-erence frame, and finally produces the relatively optimal sampling scheme by using the discrete decision-making function (built by this paper) and comparing them in combination with the dia-grams. According to the test result in the survey of the arable land using remotely sensed data, the Sandwich model, while applied in the survey of the thin-feature and cultivated land areas with aerial photos, can better realize the goal of the best balance between investment and accuracy. With this case and other cases, it is shown that the optimal decision-making model of spatial sampling is a good choice in the survey of the farm areas using remote sensing, with its distin-guished benefit of higher precision at less cost or vice versa. In order to extensively apply the model in the surveys of natural resources, including arable farm areas, this paper proposes the prototype of development using the component technology, that could considerably improve the analysis efficiency by insetting program components within the software environment of GIS and RS. 关 键 词：Keywords: SURVEY of countrywide LAND with RS data, SPATIAL sampling, optimal decision-making, model of SPATIAL sampling.