Study On Extraction Method Of Urban Built-up Area Based On Step Thinking And Multi-Source Data

Cities are the advanced products evolved in the development of human society and are closely related to human life.The area of urban built-up is an important indicator of urbanization and it provides foundational information for urban studies and meets the requirements for urban pattern and urban spatial structure research.However,the existing methods to obtain the urban built-up areas of cities based on single-source data have limited accuracy due to the accuracy and spectral confusion.Therefore,the interpretations of urban social and economic activities are not clear enough,there are great limitations of these methods.Different types of data have their own unique advantages and deficiencies in explaining urban social economic activities,such as nighttime light data can effectively extract the boundaries of the built-up area of cities,and the change of light intensity can characterize the changes of human socio-economic intensity,but it cannot distinguish the non-built-up area in the urban built-up areas and the built-up area of low light and no light area due to its low spatial resolution and light spillover.The Landsat data can accurately show the shape of urban built-up area.However,due to the spectral confusion of the urban built-up area and non-urban built-up area,it is hard to separate the urban built-up area and non-urban built-up area.As a kind of big data,POI has been used in urban studies.However,the numbers of POI in different grades of cities is quite different,moreover,POI data represent urban entities by points which decide it is difficult to show the extent of urban built-up area accurately.However,the combination of multi-source data can overcome these shortcomings and improve the accurately of-extracting urban built-up area.To remedy this,this paper proposes a method to extract the boundaries of urban built-up areas based on step-by-step thinking and multi-source data(NPP/NIIRS nighttime light data,Point of Interest data,Landsat8 OLI data).Among them,NPPANIIRS nighttime light data,permanent population data and population urbanization rate data are used to extract the peripheral boundary of a city to determine the approximate scope of urban built-up area;POI data is used to extract the urban built-up area with low light or no light areas,with the urban built-up area extract from NPP/VIIRS nighttime light data to form the final peripheral urban built-up area;Landsat8 OLI data are used to extract the non-construction land within urban built-up areas to obtain the final urban built-up area.The method was applied to the extract the urban built-up area of three cities(Beijing,Wuhan and Fuzhou)of different economic development level and urbanization level,then,the extraction results were compared with Google Earth to verify the extraction accuracy,the result showed that the accuracy is more than 95%,which proves that this method can extract urban built-up area more accurately than ordinary methods and applicable to cities of different levels.This method has certain scientific value.