Development of a methodology to map railway lines and surrounding land use using UAVs

Abstract

High accurate railway maps and terrain information (Digital Elevation Models) is a major concern for future railway constructions and railway lines development.The mapping of railway line using ground based surveying techniques istime consuming and problematic.Unmanned Aerial Vehicles (UAV) technology has revolutionized the aerial photogrammetric mapping due to its low cost and high spatial resolution. It enables mapping the land use with greater accuracy in both 2D and 3D. The “DJI Phantom 4“ drone was selected as the UAV platform to acquire image data. In this study, we have developed a fully automated and highly accurate engineering approach for detecting land use and railway line, which is based on textural information from orthophoto and elevation information (Digital Surface Models)obtained from the drone.The Pix4D software was used to develop the orthophoto and a Digital Surface Model (DSM) and the DSM was validated by using the ground control points.The rule sets knowledge-based classification method in object oriented classification was used to classify the land use and railway with the use of “eCognition“ software. Finally, the results were compared with digitize land use layer to validate the results, and obtained overall accuracy of 90.15%.