Identification method of functional areas along pipelines based on multi-source data

Traditionally, the functional areas along pipelines are identified by manual on-site survey, which has limited efficiency of the identification under the influence of the personnel experience and proficiency of identification workers. In view of this, the deep learning method was used to extract the buildings in high consequence areas based on the high-resolution remote sensing images. Meanwhile, comparative experiment was conducted between the improved model and the existing mature semantic segmentation model based on the public dataset. The results show that the building extraction network proposed in this paper has higher building interpretation accuracy. The density peak clustering analysis method was used in combination with the points of interest (POI) and the open street map (OSM) data to calculate the frequency density of various types of data, so as to identify the functional areas along pipelines. In addition, this method was applied to identify the functional areas in the high consequence of a West-East Gas Pipeline, with six types of functional areas identified, including the residential areas, public service areas, industrial areas, commercial areas, traffic facility areas, and agricultural areas. Generally, the proposed method could accurately and quickly identify the densely-populated places, thus further improving the identification accuracy and efficiency of high consequence areas.