| In recent years, satellite and remote sensing technology as the advanced spaceexploration technology obtained the rapid development，various satellite imagesappear constantly, providing fertile remote sensing image data sources. Thanks forits high resolution, true color, containing abundant geometric information, highresolution remote sensing data gradually become new favorite in the field of remotesensing data study and more application. A large number of studies have shown thatcompared with the traditional classification method based on pixels, the extractinformation classification of high resolution satellite images, and the object-orientedimage classification techniques has obvious technical advantage and higher precision.Basing on spectral information, fully considering the geometry and texturecharacteristics of the image object information, using technologies such as imagesegmentation, object-oriented classification technology could be used to categorizethe same kind of like pixel into a collection, and finally obtain the target informationin the form of video object.In this article, basing on the correlation function and parameters of ERDASObjective module, study has been carried out on object-oriented image informationextraction technology used in Quickbird satellite images of the main urban area ofShishi City（Fujian Province）and data for urban green space plaques revealed. Thenthe author performed a quantitative calculation of a variety of green space landscapeindex. Landscape pattern analysis method is applied to the situation of urban greenspace of Shishi city urban area, the green patch size, structure and landscape index analysis and evaluation. Through research proves that the use of object-orientedremote sensing image information extraction technology for the high precision ofimage information can be very good application in all kinds of the urban green spacelandscape analysis on the basis of the spatial data such as area of research.Among many kinds of object-oriented remote sensing image informationextraction software, ERDAS Objective modules is rarely used in advanced researchand application because of its complicated function. But there is no denying thatERDAS Objective module with its unique processing method and the workflowpattern could provide the object level of image processing and multi-level multi-typesachievement. It really fulfills the target of the object-oriented image classification andinformation extraction. Based on the research on ERDAS Objective detailed functionand parameter module of a variety of functions, and according to the features of greenspace information of vegetation and the actual situation in the study area of repeatedextraction experiments, in this article I ultimately made the conclusion that theoptimal segmentation threshold of could be65.5. I believe that extraction workflowcould be formed by the probability filter, size filter, expansion, corrosion, clusteringand smooth, and all other applicable functions. The producers’ precision and users’precision of extraction of green patches in the study area data is as high as98%and96.1%, respectively, which is completely meet in application of study area of urbangreen space landscape analysis.According to the extraction of greenbelt information, from the prospective of thebasic index, green patch size and statistical calculation of urban green space landscapeindicators, this article comprehensively analyses the main urban area of Shishi city ofall kinds of green space landscape, including Fengli street, Hubin street, the town ofLingxiu and the town of Baogai. The results showed that per capita and overall levelof green land coverage is better in Shishi urban greenbelt area. A large number ofdistributed discrete residential green spaces caused the great fragmentation of urbangreen patch in downtown. The aggregation and richness of green area is relativelygood. Urban green corridor network improves the integration of t green space plaquesin urban area. The green space distribution pattern of urban area in Shishi city mainly consist of large park green space, the green corridor of residential green spacenetwork, along with discrete distribution. With the rapid development of Shishi cityeconomy, the government pushing forward the urbanization process, the number andintegrity of agricultural land and natural forest land is gradually decreased due toerosion of urban construction in the whole city Shishi city and the surrounding area.Producing green area, mainly agricultural land is gradually reduced. Landscape forestlands in Peri-urban area become important green space. With the development ofurban traffic, its influence in regulating the urban ecological environment and inmaintain the city’s image as “livable city” will also gradually be presented. Specialattention should be drawn to strengthen the protection and exploration. |
PDF Full Text Request