Study On The Effect Of Different Data Sources On Urban Green Space Landscape Pattern Evaluation

In this paper, the impact of different resolution remote sensing data source on the landscape pattern analysis of Suzhou City has been explored based on the built-up area of Suzhou, combined with the landscape pattern analysis and3s technology, which has provided as the scientific basis for future studies of landscape pattern of urban green space system. Meanwhile the overall pattern and gradient feature of Suzhou City green landscape has also been analyzed, in order to help the construction of Suzhou urban green space system.In the paper, three different phases of resolution images has been selected, which are Landsat TM image dated by September21,2010; ALOS image dated by May3,2010, and IKONOS image dated by April27,2011. Knowledge-based classification method with high-precision has been used on three data sources for urban land classification; the high-precision green information has been extracted through the software and manual methods. Subsequently, thematic maps of the Suzhou city green space system corresponding to three data sources have been produced by means of overlay analysis. Due to different data sources, the impact of urban green space landscape pattern has been studied from the global and local two aspects. The main target of the study has been completed, and the main results are as follows:(1) Traditional supervised classification has been compared with knowledge-based classification method. The conclusion is that the supervised classification does not apply to urban land use classification system, while knowledge-based classification method meets the requirements.(2) Size effect of different data sources has been studied, and the appropriate particle size and magnitude of each data source has been selected to evaluate the landscape pattern. The appropriate particle size of the Landsat TM is1m≥m-20m, with suitable amplitude≥4km; ALOS is suitable to granularity of lm-5m, with suitable amplitude≥1km;IKONOS applies to the granularity of1m-2m, with suitable amplitude≥1km.(3) In the study of the overall urban green space landscape pattern, a relatively high spatial resolution images need to be chosen for such studies concerning the fragmentation of urban green space, such as ALOS and IKONOS, When studying the diversity of urban green space system, the advantage of the landscape or aggregation, relatively low resolution image of can be chosen, such as TM.When analyzing the gradient of urban green space system, images with high or medium resolution should be selected. While images with low resolution, like TM, are not suitable for feature analysis of urban green space concerning green space type, etc.(4) The overall landscape pattern characteristics of urban green space system of Suzhou, has been summarized as:unreasonable green plaque composition, with highlight advantages of type; relatively high level of landscape fragmentation, and complex shape; higher landscape heterogeneity and aggregation.The gradient feature on the landscape level along the north and south of the People’s Road of Suzhou has been expressed as:the green landscape of the central urban patch density is higher than the suburbs, and south of district higher than the north; the landscape diversity of suburbs and the city center is lower than the middle of the city; and the aggregation of city center is relatively low.