Font Size: a A A

Information Extraction And Risk Evaluation Of Vegetation Destruction In Natural Forest Region Based On3S Technologys

Posted on:2013-01-12Degree:MasterType:Thesis
Country:ChinaCandidate:J LuoFull Text:PDF
GTID:2233330395978766Subject:Forest management
Abstract/Summary:PDF Full Text Request
Natural forest, having distinct effect on improving ecological environment, regulating climate, conserving water and soil, keeping ecological balance and protecting biological diversity, is the main part of forest resources in our country and it brings much ecological and environmental benefits. After Natural Forest Protection Project is carried out, banning deforestation measures is also taken in natural forest areas, which fundamentally controls the situation of human cutting on serious deforestation of natural forest. Besides, there still exists many natural and human factors, which will destroy larger areas of natural forest resources.As research area Mu cheng Ditch, the core region of Jiajin Mountain National Forest Park, using QuickBird high spatial resolution remote sensing images, through digital terrain analysis based on DEM, combined with the comprehensive evaluations of information value model on the risk of all kinds of damage on natural forests, the feasibility and validity of information extraction using object-oriented classification methods for natural forest destruction was explored in the paper. It provides systematic researches on the destruction of natural forest with a "point" of the model, provides technical support for the protection measures of natural forest resources and explores protection and improvement of ecological environment in our southwest forest region. The main conclusions of the research are as follows:(1) Based on the summary and analysis of image features of landslide, roads and habitations, by using object-oriented classification methods the three destruction types in QuickBird image were automatically extracted and precision evaluation was made. The results showed that the position precision of the landslide, roads, habitations of was90.91%,87.79%and86.86%and shape precision was92.69%,90.58%and86.00%.(2) The data of the risk evaluation factors in the destructive areas was classified, statistically analyzed and its information value was calculated. It showed that destruction of the landslide was mostly likely to occur in bare land, where the slope was bigger than50°,the density of gully was more than7km/km2, and was100m from the road. Grazing region was mostly likely to occur in grassland area, where the slope was less than20° and was500m away from the habitations. The wind throw destruction was mostly likely to occur in forest region, where elevation was from3500to3800m of the northwest slope and the canopy density was less than0.5. The snow pressure destruction was mostly likely to occur in areas where elevation was more than3500m in the Northern Slope and tree species in the area were spruce or fir.(3) Using information method, the grading evaluation results of risk of damage and extracted information of known damage were analyzed. It showed that, with the increase of the risk level, damage information contained in the grading area increased and87.93%known damage information was found in severe and very severe areas, which showed that through information method, the risk evaluation results and the actual damage in the research area matched and the division of risk grades was reasonable.(4) Different risk levels in the research area were statistically analyzed. The results showed that the largest distribution area is moderate danger district in the research area. The total area of moderate danger district was1812.3218hm2, accounting for33.48%of the research area. The total areas of mild and a very mild danger districts was2145.46hm2, accounting for39.63%of the research area. The total areas of severe and very severe hazardous districts was1455.4533, accounting for26.89%of the research area.(5) From the disposition of extremely severe danger zone in the research area, these areas mainly occurred in both sides of the road and it could be speculated that the main factor inducing destruction was the roads in the research area.
Keywords/Search Tags:3S techniques, object-oriented classification methods, DEM, informationmethods, risk evaluation
PDF Full Text Request
Related items