基于HJ-1A星HSI高光谱数据的南疆盆地主栽经济林树种识别研究Based on HJ - 1 a star HSI hyperspectral data of southern injiang basins main economic forest tree species identification study.docxVIP

Classification Classification for Main Fruit Trees Species in Southern Xinjiang Basin Based on the HSI Hyperspectral Data of HJ-1A Satellite Abstract Recent decades．Southern Xinjiang Tarim Basin has become the main producing areas of fruit characteristics．Xinj iang rural economic development and rural incomes leading industry, ground．based survey of traditional forestry survey method，time—consuming，long and labor-intensive．high cost and cycle；Multispectral remote sensing to supplement resources investigation indistinguishable species．Can not meet the needs of southern fruit industry development project inforillation management．Due to the advantages of a narrow-band． multi．channel．into one image and the spectral．hyperspectral data Can accurately detect the subtle spectral feature types have differences． For the hyperspectral remote sensing．related issues in forest tree species identification， Study on hyperspectral data HJ．1A satellite(Hyper-Spectral Imager,abbreviated HSI)as data source．and a large area of Akesu region the W．ensu county in around the artificial planting of apple(Malus pumila Mill)，walnuts(Juglans regia Linn)，jujube(Ziziphus zizyphus Mill) three fruit for the study,using the different denoising and dimension reduction way to process images．based on the different denoising and dimension reduction way to process images．To carry out recognition of forest tree species based on the feature space and spectral space．，expand the application areas of domestic satellite data,with a view to the southern border of Main forest tree species identification of Remote Sensing and management of forest resource information and provide scientific basis．The obtained results are as follows： (1)Identification of forest tree species in South cultivars based band selection shows that： Different band selection method for screening out sensitive band concentrated in the red and near-infrared wavelength range，and of forest tree species have good separability；Sensitive band com