The Study Of Rapidly Acquiring Forest Cover Information Technique For China

In order the ecological construction of the forestry development strategy can be put into practice, in the meantime can also promote the modernization of forestry construction, accelerate land afforest and strengthen resource protection, the forestry sector has made the realization of modern management for forestry resource as one of their vital tasks. With the development of 3S technology and under the help of the remote sensing to real time investigate and monitor the forestry resources, the forestry sectors can scientifically manage and utilize the forestry resources sustainably, all these can provide scientific and effective data support for government departments.Based on the Vegetation Regionalization Map of China, the research divided the whole nation into 8 regions to extract the forest cover information. The model algorithm and technical method researches are mainly focused on the Cold temperate zone with coniferous forests region with high forest coverage and variety vegetation types.(1)Based on the objective law of factors such as the terrain, temperature and rainfall, divided the nation into 8 regions to extract forestry coverage information..(2) In order to rapidly obtain separable and representative training samples on large area remote sensing images, this study based on the macroscopical character of remote sensing images, referring to the 1:1000000 vegetation map and the China land cover map named as WESTDC, combining the unsupervised classification results of experimental region from the 2001 MODIS time series NDVI images. The spatial features of vector and raster data were also used and associated the unsupervised classification results as the random sample properties. Based on the unsupervised classification types and the random point ratio of different classification types to realize the sample purification, then compared the difference of the standard deviation and the separability index before and after the sample purification of different vegetable types samples, the points which have relatively large contribution to the supervised classification were saved. The results showed that the land cover spatial distribution in survey area was consistent with the WESTDC map, the overall accuracy of main types was 84.82%. Then used the purred samples and applied with maximun likelihood classification method., the overall accuracy was increased 25.62% than that from initial samples. It means the research offering a valid sampling method for remote sensing images of large area and reducing time losses, manpower and material costs.(3) The research on information extraction technology of vegetation type in large area based on multiple classifier combination.The research used the combination rules of evidence theory to extract the vegetation types in large area. Based on the classification system design evidence theory and with the classification rules of the maximum likelihood, minimum distance, and sequential SAM to get the classification results. The results will be normalized processed under the IDL program algorithm, and would get mBPA of three different classification rules, to avoid the zadeh paradox which caused by the image noise, the research set up the superset concept set. Combined the mBPA of different evidence sources and calculated the combined belief function. The overall accuracy of maximum likelihood, minimum distance, and sequential SAM are 78.14%,72.56%, 76.19%, respectively, while the overall accuracy of the combination rules of evidence theory is 80.84%.(4) The extraction of national forest cover information.Based on the classification process of rapid sampling and combine classifier, the research used remote sensing to classify the nation’s vegetation species, the 8 different vegetation zone are the Cold temperate zone with coniferous forests region, Temperate zone with coniferous and deciduous broadleaf mixed forests region, Warm temperate zone with broadleaf deciduous forests region, Subtropical zone with broadleaf evergreen forests region, Tropical zone with monsoon forests and rain forests region, Temperate zone with grass region, Temperate zone with desert region, Tibet Plateau high-cole vegetation region, respectively.The extraction results are in accordance with the Vegetation Regionalization Map of China. The results showed the national arbor coverage area is 1500660km2, the coverage rate is 15.69%, the bush coverage area is 844950km2, the coverage rate is 8.83%. Based on the Vegetation Regionalization Map of China, the national arbor coverage area is 1499000 km2, the coverage rate is 15.67%, the shrub coverage area is 891400 km2, the coverage rate is 9.32%. The tree and shrub coverages are different about 0.02% and 0.49%. The tree coverage decreases by 0.47% than that in the 6th national forest resources inventory...