Millimeter accurate multi-view stereopsis reconstructions of railway profiles in Southern Africa
This dataset forms part of the thesis titled : Railway conditioning monitoring using a hybrid multi-view stereopsis and real-time kinematic geolocation pipeline for dense reconstruction:
- The "MVSNet.zip" archive contains the associated Spooroog software (Python) and trained neural networks that were developed to post-process the datasets, as described in Chapter 6 of the thesis.
- The "Experiments.zip" archive contains measurement and experimental data as described in Chapter 7 of the thesis, obtained from both the University of Pretoria's Test Track faciilty (LiDAR (Hovermap), Spooroog, digital level, UAV, structured light scanner (EinScan), geometry measurement system (KRAB) and real-time kinematic antenna surface mapping (RTK-ASM)) as well as in Ermelo, KwaZulu-Natal (Hovermap LiDAR and Spooroog). These measurements of railway profiles are subject to a variety of illumination and material conditions, represent both existing and novel measurement and instrumentation techniques investigated.
- The "Analysed results.zip" contains the processed data as presented in Chapter 8 of the thesis. These are subdivided into individual sub-chapters corresponding to the thesis. This also includes the Python scripts used to generate the graphs and composite visuals.
Each file is a compressed archive (zip) that can be accessed using an archive manager after downloading.