Multi-scale Urban Land Use/Cover Change Detection And Temporal And Spatial Analysis

Land use and land cover(LULC) change in urban area is an important indicator to quantify not only the characteristics and extent of urban development but also lots of effective factor of ecological environment in cities. Therefore, detecting land use/cover change(LUCC), rapid and accurate, is the basis of many urban-related researches. Due to the densely populated urban areas, complex human activity, composition and changes of land use/cover is also very complicated. And urban LUCC detection and feature extraction has been a hot field for remote sensing applications, for some time now. But for the high heterogeneity in urban area, scale issues must be considered in LULC researches. Spatial resolution of the data scale and LULC information of the operational scale on the behalf of urban LULC information acquisition and LULC spatial and temporal characteristics researches become an important connection between studies, especially the image resolution. But the function between spatial resolution and change detection or the representative of the data resolution for urban LULC research is uncertain. Therefore, studying change detection methods for the changing spatial resolution is important to accurately extract urban LULC information. To clear the suitable geographic attributes of scale of LULC data is also very important to promote the urban LULC research and to ease various urban environmental problems.We take Shanghai as our research area and build a multi-method to detection changes in our study area from data scale and operational scale. We acquire the 2005 and the 2010 LULC information in Shanghai and discuses the suitableness and effective for different resolution data to get the LULC information. And then we use the landscape ecological-method to quantify LULC patterns spatial variations. And the main contents and conclusions are as follows:a) We offer a multi-method change detection technology. The method can effectively eliminate pseudo change information, improving urban LULC classification accuracy;b) We perform change detection under two different resolution(2.5m and 30m). We find out that the erroneous detection under 30 m is mainly caused by material difference vegetation phenology and the erroneous detection under 2.5m is mainly caused by the angel of sensor. All of this is because of the great difference of information capacity between the two resolution.c) The analysis of spatial and temporal distribution of urban LULC of different operational scale shows that the trend between different scale is different.d) It is a fitting data scale that can reveal the phenomenon under the research scale.