The Rresearch Of Urban Green Space Abstraction And Classification Based On QuickBird Image

Urban green space, as the primary producer of urban ecological system and the controller of urban ecological balance, plays an important role in urban environment regulation. Therefore, an increasing number of cities have begun to do the survey to get the latest basic material concerning urban green space in order to provide the better planning basis and management decision.he paper uses the current commercial satellite-Quickbird for study and the study area is Dongguan city in Guangdong province. The study process is as follows:first, the remote sensing data is pre-processed, and then, the vegetation index model, which is quickly acquired, easy to be operated, is used to extract the urban green space. Because the model is extracted by the spectrum information, it is easily to be disturbed of the information concerning "same object with different spectrums" and "same spectrum with different objects", which affects the accuracy and speed of the information extraction. It is found that the information concerning "same spectrum with different objects" is mainly the blue roofs, white architectures and farmland through surface features identification, while the information regarding "same object with different spectrums" is mainly street trees. After that, these types of surface features are compared with the real vegetation information of the urban green space in the form of spectrum sampling comparison analysis, and the vegetation extraction spectrum model is established to avoid extraction of the blue roofs and white architectures at the same time and eliminate the phenomenon of "same spectrum with different objects". Besides, the farmland extraction method of object-oriented classification technology is used to lessen the influence of the "same spectrum with different objects" and it is verified feasible by experiments. Due to the insignificant difference of the information concerning the spectrum and shape of green space in the residential area or affiliated to units, the computer can not implement the automatic classification. Therefore, the knowledge-based classification means is applied based on the special consideration of the social attributes owned by the urban green space. In addition, combined with the spatial analysis function of GIS, the urban land use map provided by the planning department is fully used to realize the final classification of urban green space. And the database containing various types of urban green space information is established for management and statistics.According to the experiment and analysis in this paper, the phenomena regarding "same object with different spectrums" and "same spectrum with different objects" have been preferably solved, with obvious improvement in accuracy. In addition, the study has greatly lowered the requirements for the knowledge reserve of the visual interpretation workers. And it also tries its best to lower the waste of human and material resources, provide the methods and technique support for quick and accurate extraction of the urban green space. Finally, the study has made the relevant departments much closer in the working contents to provide effective approaches for the management and decision-making of the urban green space.