Land use classification in construction areas based on volunteered geographic information

In the process of land use management, the relative low construction land use efficiency may lead to more other types of land into construction land, and thus will also affect the amount of cultivated land, while reducing the ecological land as well. According to the current land use classification standard (GB/T 21010-2007), the construction land can be divided into business service land, residential land, public management and public service land, and storage land for industry, etc. To enhance the utilization efficiency of construction land, this research aims to develop a better way to divide the space distribution of different current land use types, which also provide both database and clues to concentrated land-use planning and monitoring. In this study, a new methodology (hierarchical grading classification method) will be proposed to solve the problems existing in the traditional division methods of construction land, and the experimental results will deliver a series of meaningful interpretations across discipline. To demonstrate this, taking the Fifth Ring of Hai Dian district, Beijing city as the research area, a variety of volunteer geographic information is determined, which including the Open Street Map (OSM), Points of Interest (POI), blogging sign data and Panoramio photos, etc. Firstly, the Open Street Map road data is used as a block boundary to divide the construction land into the different hierarchy of land parcels. Point of interest, its essence is the abstract expression of geographical entities. Since the Point of Interest and the divided land parcels share the consistent feature attributes, it is possible to use the different grade of POI to assign attribute to the different hierarchy of land parcels, and then the final results of the multi-layers will be combined together to do analysis (overlay, merge) to determine the construction land use type in this region. Finally, the confusion matrix is generated to compare the results among the Google street view, fieldwork and urban planning map. The accuracy rate of commercial and business facilities, industrial and warehouse, residential, administration and public services, street and transportation, and other construction land are 94.7%, 69.2%, 81.4%, 75.0%, 96.7% and 74.7% respectively. Furthermore, the kappa indices of classification is 0.83, showing that in this study, both the adopted data and the newly proposed method used in the process of classification of construction land are feasible, and the new method will have significant impact on the process of division construction land.