Monitoring Method In Returning Cropland To Forestland Project And Land Use Classification Based On ALOS

In order to improve the rapidly deteriorating ecological environment in China, China implements six forestry projects, natural forest protection, returning cropland to forest project, sandstorm source control projects in Beijing and Tianjin, three north and the Changjiang River protect projects and other shelterbelt construction projects, wildlife protection and Nature Reserve Construction Project, focusing on fast growing timber base area construction are included. The scope of Beijing and Tianjin sandstorm source project, including five provinces,Beijing, Tianjin, Hebei, Shanxi, Inner Mongolia(cities, counties), including 75 counties. Select a significantly summer vegetation remote sensing image which has a higher resolution as remote sensing information source is very useful.In this paper, set ALOS satellite remote sensing data as the data source, the supervised classification by computer and human-computer interactive interpretation of the actual project area is used to extract project area, but the cost of the investigation of all vegetation polygons is very costly and time-consuming. Though unequal sampling method (PPS), remote sensing interpretation given polygon area/polygon area of the Project as a secondary factor, according to 30 projects area and forest area measured by GPS, estimate saved project of the total afforestation area, and compare the two extraction area. The results show that the analysis accuracy is 86.2%.It could help to grasp the implementation of the project progress and quality,which is very important.Jiuliancheng Town has a strong heterogeneity in ground, and land use is fuzzy.The method of extract model in remote sensing images of land use to obtain accurate information which has a great significance. In this paper,through establishing information model of various types, obtainning accurate land use for pastoral areas and providing effective reference for complex spectral features. According to the basic components, linear spectral mixture analysis model is used to decomposition remote sensing images of ALOS. The reseach area is divided into four types,including forest land use, cultivated land, saline land and grassland,without water bodies, through the mask to reduce the risk of linear spectral mixture model decomposition. The decomposition of ALOS, according to root mean square error statistics,the average of RMS is 0.053 and variance is 0.018, distinct the four types of land use clearly. According to research on monitoring returning cropland to forestland Project and distinct the type of land use, the scientific management of agriculture and animal husbandry ecotone, provide effective decision support.