Classification Of Forest Resource And Estimation Of Forest Canopy Closure In Kunlun Juniperus Nature Reserve， Xinjiang

Classification of forest resources and forest canopy closure estimation is veryimportant in forestry production that especially in the Forest Resource Inventory andsub-species statistical area of forest canopy closure and volume survey factor.Therefore, the classification of forest resources in forestry should be the moreaccurate the better. The estimates of forest canopy closure and inversion is as accurateas possible. Conventional forest resource classification and forest canopy closureinvestigation relies mainly on the artificial field investigation or the use of large-scaleaerial photographs. These two methods has its deficiencies are the labor intensity ofthe artificial field investigation is too large and the use of large-scale aerialphotograph of the survey cost is too high. The rapid development of satellite remotesensing technology provides a new way for the classification of forest resources andforest canopy closure estimation.The Kunlun Juniperus Nature Reserve in the study area mainly carried out by theLandsat-7ETM+and SPOT-5images of forest resource classification study. And theapplication of GE images of forest canopy closure estimation study. The forestresource classification accuracy and forest canopy closure estimation accuracy wasanalyzed and evaluated. The main conclusions of the study are as follows:（1）Classification of forest resources. Firstly, this paper using unsupervisedclassification method for a preliminary classification of forest resources in the studyarea. The vegetation coverage area and vegetation cover were separated. Secondly, theforest resources of the study area to supplement the human-computer interactionclassification and classification. The grassland and forest areas in vegetation coverwere separated.（2）Species identification of forest resources. The combination of RS and GISand the topographic factors （elevation, aspect, slope） in study area with theclassification of vector overlay. The use of hierarchical classification method toachieve a classification to identify the main species of forest resources.（3）According to the latest results of the classification in forest resources. Theforest coverage of Kunlun Juniperus Nature Reserve is13.3%. The dominant treespecies of forest vegetation to tree forest are Picea schrenkiana. and Juniperusjarkendensis Kom.. The dominant tree species of shrub ghost are Caragana jubata.,Rosa., Caragana sinica., Lonicera japonica., Nitraria tangutorum., Hippophaerhamnoides., Cotoneaster acutifolius. Verify the accuracy and the overall averageclassification accuracy is88.65%.（4）Forest canopy density estimation based on remote sensing image textureinformation. By analyzing the texture of the GE satellite image. There are eighttexture parameters extracted from the Mean, Variance, Homogeneity, Dissimilarity,Second Moment, Contrast, Entropy and Correlation. These eight texture factorscorrelation analysis and principal component analysis will be screened three texturefactors （Mean, Second Moment, Entropy） for canopy density estimation modeling.（5）Multiple linear regression of the three factors regression equation of the four estimation models significant F test, T test and the standard regression coefficientestimation error test. And get contains Mean（M）, Second Moment（S）, Entropy（E）. The model fit best in. The linear regression equation is:Y=0.557-0.001*M+2.098*S-0.302*EVerify the accuracy and canopy density estimation model. The averageestimation accuracy is86.4%.（6）Estimation model of remote sensing inversion in the Kunlun JuniperusNature Reserve forest canopy density and forest canopy density the most widelydistributed intermediate. That’s followed by sparse level and the least dense. Finally,the forest canopy density is distributed at different elevations and slope.