The Study On Accessibility Of Urban Green Space And Analysis Of Landscape Pattern Based On Domestic GF-1 Satellite Image

Urban green space is an important part of urban landscape,it plays important roles in improving the urban ecological environment,providing various social benefits and services for the city.Researching the accessibility of urban green space and analysising the pattern of the landscape spatial,and discussing the impact on the accessibility of green space pattern,can evaluate the rationality and fairness of the urban green space layout,it plays significant roles in improving the service level of urban green space and the living environment.This paper is Supported by GIS and remote sensing technology.It studies Donghu district and Xihu district in Nanchang and uses GF-1 image in the study area as the main data source.It discusses the applicability of different fusion methods for GF-1 image which is based on the pretreatment of the original GF-1 image.Then using object-oriented classification method which is based on rule set to extract information of land use for GF-1 image in the study area.On this basis,using the weighted distance cost method for the classification of urban green space accessibility in study area,and accessibility of green space in study area to make the evaluation for the space distribution and the service level.Finally the green space landscape in study area are analyzed in plaque composition,shape index,patch fragmentation and overall structure analysis.It explores the influence of green space landscape pattern of the study area on accessibility,and the suggestions for green space planning and construction in the study area are put forward.The main research results are as follows:(1)Specify the different image fusion methods for the different applicability of GF-1.For GF-1 image in study area,the fusion effect of Brovey fusion method is poorer,HPF fusion method has the smallest spectral distortion,and the spectral information is kept the best;PCA fusion method in strengthening of spatial information and reserving spectral information both have the ability to moderate;NNDiffuse fusion method in the spectral information is not the best to maintain,but it has the best separability of things,the most layered objects,and the image contains the richest information.(2)The best segmentation scale and the merging scale of the image are obtained by a large number of experiments.The large numbers of objects samples in study area are analyzed,and then setting up the extraction rules of land use types of GF-1 image,so as to realize the land use information extraction which is based on object-oriented.(3)The accessibility level of green space in study area is determined.The distribution of green accessibility in both study areas is not balanced,but the overall accessibility is good.Most residents in the study area live in areas with accessibility of primary and secondary levels,and the level of green service of the two study areas is almost the same.(4)The spatial pattern of green space in study area is analyzed.The landscape structure of the two study areas is single and the landscape diversity is not high.The uniformity of the spatial distribution of the green space needs to be further optimized,and the overall landscape aggregation degree is high.The shape of the patch of green space is heavily influenced by human intervention,and it destroys the natural ecological succession.(5)The influence of landscape spatial pattern on accessibility is summarized.Overall,The larger the area of green patches,the greater the number of plaques,the more balanced the spatial distribution of landscape space,the better the accessibility of urban green space.