Data fusion of extremely high resolution aerial imagery and LiDAR data for automated railroad centre line reconstruction

The quality of remotely sensed data in regards of accuracy and resolution has considerably improved in recent years. Very small objects are detectable by means of imaging and laser scanning, yet there are only few studies to use such data for large scale mapping of railroad infrastructure.In this paper, an approach is presented that integrates extremely high resolution ortho-imagery and dense airborne laser scanning point clouds. These data sets are used to reconstruct railroad track centre lines. A feature level data fusion is carried out in order to combine the advantages of both data sets and to achieve a maximum of accuracy and completeness.The workflow consists of three successive processing steps. First, object-based image analysis is used to derive a railroad track mask from ortho-imagery. This spatial location information is then combined with the height information to classify the laser points. Lastly, the location of railroad track centre lines from these classified points were approximated using a feature extraction method based on an adapted random sample consensus algorithm. This workflow is tested on two railroad sections and was found to deliver very accurate results in a quickly and highly automated manner.