森林资源信息对经济发展和环境保护等具有重要意义。但是森林资源调查不仅需要大量财力，而且调查周期长，观测样地定位和复位精度低，调查成果质量难以满足要求。蓄积量调查是一个不可见的被估测因子，无法直接利用高空间分辨率卫星影像。如何根据监测区域少量地面调查样地资料及高空间分辨率遥感图像，借助空间相关性理论和非线性理论，建立以样地为单位的森林蓄积量估测方程具有重要的理论研究价值和经济意义。 随着3S技术的发展，基于传统抽样方法和3S技术相结合的空间对地抽样方法为监测区域变量估算提供了新的思路。利用遥感影像获取的统计抽样中总体的先验知识，可以提高抽样调查的精度，结合地面调查，可以减少野外样本量，降低调查成本。空间相关性技术应用于遥感领域，可以指导监测区域样地选取，并可通过空间相关性分析技术对样地的布设模型进行评价。 本文首先根据思茅地区监测区域地面调查样地的林业区域变量信息，包括蓄积量、郁闭度和样地株数，利用地统计学理论分析了当地森林资源空间异质性、方向异性、空间分布格局以及空间相关尺度，为林木蓄积量的估算方程进而为样地的选取提供了依据。然后本文根据经典抽样理论和空间抽样理论，研究了空间对地抽样不同的方法，以及抽样数量对样木蓄积量估计精度的影响，研究表明利用遥感图像与部分样地调查数据相结合获取的先验知识，可以较大幅度提高估算精度；同时利用精度-样本图，为最优化决策提供依据。

Forest resources information is very important for economic development and environmental protection. But investigation of forest resources needs not only more money, but also more time, meantime, the precision of positioning and re-positioning of the observed sample places is low. So the investigation result and quality can’t meet requirement. Stock volume which is an important factor in forest resources investigation can’t be obtained directly by remote sensing image. How to build the forest stock volume estimation equation based on the exampling region is very important on theoretical research and economic. This not only needs sample places information and remote sensing image, but also needs spatial correlation theory and nonlinear theory. With the development of 3S, it provides a new way for us to estimate the variables of observed sample region by combination of traditional sampling method and spatial sampling method based on 3S technology. Making use of the knowledge of population of statistical sampling obtained from remote sensing image, we can increase the precision of investigation and sampling, integrating the manual survey we can decrease the quantity of the sample places and can reduce the cost of investigation on sample places. By applying the spatial correlation technology on the remote sensing, the selection of sample places can be guided, and also the distribution model of sample places can be evaluated by spatial correlation analysis technology. In this paper, we researched the area of Simao in YunNan province. Firstly, according to the forest area variables information of investigated sample places in observed region in Simao, including forest stock volume, forest canopy density, and quantity of trees in sample place, making use of Geostatistics, we analyzed the spatial variability law, direction variability law, spatial distribution characteristics, and spatial correlation scale of the forest resources in Simao, which is the basis for estimation equation of forest stock volume, and therefore for sample places selection. Secondly, according to the classic sampling theory and spatial sampling theory, different methods of sampling from space to ground, and the influences of sample quantity to forest stock volume estimation precision were studied. The study showed that the estimation precision is increased significantly by making use of knowledge obtained from combination of remote sensing image and some sample places investigation data, at the same time, the chart of precision-sample arbor which can be achieved by spatial analysis is the precondition for the optimal decision.