Based On Attribute Of Sample Places In Observed Region Of Spatial Correlation Analysis

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.