Assessing Method Research On Forest Survey Factor Based On IKONOS Image And DGPS

Since the appearance of satellite remote sensing, forest operators have developed series of research on forest resource survey application in order to apply this new techniques into large-area forest resource survey. With the in-depth study and integrity of technique, the application range of RS can be divided into several parts: forest resource survey and plan, dynamic monitoring and analysis, natural disaster monitoring and forecast, assessment of disaster, quantitative estimation and so on. However, because of the restrict of resolution, forest quantitative research concentrated primarily on forest stock volume estimation, while few research on forest land survey factors estimation such as mean diameter, mean tree height and so on.In 1999, the IKONOS satellite of USA launched successfully, which indicated the coming of a new remote sensing era. Then many countries put great emphasize on IKONOS Image. At present, IKONOS has applied in some special fields, including forest land use and city cadastral management. There are several research on isolated trees species classification and estimation of tree canopy diameter based on IKONOS in other countries. All the research showed that high resolution remote sensing can provide a new approach to forest quantitative research.In this paper, WeiJiaCun Centre, one part of Beijing West Mountain Forestry was experienced using IKONOS as the main data source. And systemic research has performed on the key techniques of resolution remote sensing of forest quantitative research. The purpose is to find the application potential of high resolution satellite for forest quantitative research and make up the deficiency of low resolution RS.Main Research Contents:1. Research on the data process method for improving DGPS data accuracy .2. Research on DGPS positioning accuracy under the tree canopy in forest land.3 .Research on accuracy analysis of RS Image orthorectifation using the Ground Control Points collected by DGPS.4. Confirm the setting of estimation variables of forest survey factors and Optimizing the key factors based on RS and GIS,5. Analysis on the effect of parameter estimation and forecast accuracy according to the select of variable factors.6. The automatic extraction method on the RS information and geography information corresponding to sample plots.7. Analysis on the Multi-Collinearity and abnormal value effect in the quantitative estimation of forest survey factors.8. Application of resist-PCA estimation in the estimation of forest survey factors. Primary conclusion and results as follows:1. Pseudo range smooth and phase smooth can effectively improve the accuracy of DGPS positioning2. According to the analysis of positioning experience data, it was concluded that DGPS is can afford to forest land positioning with the allowed accuracy. The mean positioning accuracy under the tree canopy is good, while the error fluctuated greatly, which should be study further.3. Using the ground control points collected by DGPS to correct the image, the corresponding image data were registered by the sample plots. Through comparing different correction methods, it was concluded that : malformation in image brought about by terrain factor is very big, and orthorectification is needed to eliminate the effect of terrain. And among the methods, second-degree polynomial orthorectification is the best one.4. Combined GIS with MATLAB, the automatic extraction of RS and geography information of sample plots was realized.5. Introduced the resist-PCA estimation method to build the estimation formula on crown density, stock volume and diameter breast height, good forecast accuracy was achieved and the deficiency of traditional LMS parameter estimation was made up.6. The estimation process software of the model was developed, including data pretreatment, weight function selection, LMS estimation, PCS estimation, resist-PCA estimation, accuracy assessment and other function models, in order to c...