Extraction of Farmland Geographic Information Using OpenStreetMap Data

Abstract

Volunteered Geographic Information (VGI) is a kind of geographic information which is created, edited, managed and maintained on the basis of common handheld GPS terminals, high resolution remote sensing images and personal spatial cognition. VGI has received extensive attention from scholars at home and abroad for its advantages of wide data coverage, strong current situation, and free use. OpenStreetMap (OSM) is simple in its way of generating and uploading data, and it has a large amount of storage data, which has become a widely used case model in VGI. Since there is no effective classification and management of data, how to extract valuable information from OSM data effectively becomes a bottleneck restricting the application of OSM data. Thematic maps of agricultural information can visually reflect the current status of agriculture in a region and have important reference value for government departments in formulating agricultural policies and economic plans. At present, thematic map production of agricultural information is mainly completed using satellite remote sensing and ground surveys, which requires higher costs. But agriculture is a low-value-added industry. How to reduce the production cost of agricultural thematic maps is an issue that needs urgent solution. In this study, Hapcheon County in South Korea is selected as the research area. First, study the threshold setting for the area of blocks, the area of buildings in the blocks, the density of road lines, the density of cores, and the density of road nodes, etc., and use the recent historical OSM data of the study area to extract the area of cropland; Secondly, the farmland area extracted from the OSM data was compared with the cultivated land area of the combined Unmanned Aerial Vehicle (UAV) and RapidEye image data to determine the optimal threshold; Finally, the farmland information based on the optimal threshold value and OSM's own agricultural geographic information overlay are integrated to realize the agricultural geographic information thematic map production based on OSM data. The research results show that setting thresholds for various impact factors, such as block size, can improve the extraction accuracy of farmland area. Based on the information of cultivated land area extracted by remote sensing image of UAV, the information of farmland based on OSM data has good accuracy (The error is about 15%). The study can achieve rapid acquisition of geographical information of farmland at various regional scales and provide data support for dynamic monitoring of farmland area and agricultural production management.