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 o...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.展开更多
针对抽样调查中抽样设计、估计量设计及方差估计等方面存在的关键理论性问题,运用数理统计方法,从抽样调查的两个主要环节,即抽样设计和抽样估计环节进行基础理论的综述研究,以S(a|¨)rndal et al.(1992)等成果中研究的抽样设计、...针对抽样调查中抽样设计、估计量设计及方差估计等方面存在的关键理论性问题,运用数理统计方法,从抽样调查的两个主要环节,即抽样设计和抽样估计环节进行基础理论的综述研究,以S(a|¨)rndal et al.(1992)等成果中研究的抽样设计、示性变量、包含概率、π估计量等核心概念为基础,并引入超总体模型这一研究工具进行模型辅助估计,最终归纳整理出一套现代抽样调查的基础理论体系,为后续更好地开展抽样调查基础理论和应用研究奠定方法论基础。这套基础理论体系具有开阔性、统一性和易于推广性等一系列优势,对于抽样调查从设计到估计的全过程起着基础性作用。展开更多
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)+1 种基金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).
文摘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.
文摘针对抽样调查中抽样设计、估计量设计及方差估计等方面存在的关键理论性问题,运用数理统计方法,从抽样调查的两个主要环节,即抽样设计和抽样估计环节进行基础理论的综述研究,以S(a|¨)rndal et al.(1992)等成果中研究的抽样设计、示性变量、包含概率、π估计量等核心概念为基础,并引入超总体模型这一研究工具进行模型辅助估计,最终归纳整理出一套现代抽样调查的基础理论体系,为后续更好地开展抽样调查基础理论和应用研究奠定方法论基础。这套基础理论体系具有开阔性、统一性和易于推广性等一系列优势,对于抽样调查从设计到估计的全过程起着基础性作用。
