RAG has been producing hard coal underground in Germany until the end of 2018 and continues to be responsible for the resulting consequences in North Rhine-Westphalia and Saarland. For almost 200 years, thousands of mine shafts and other cavities have been developed underground, many of them at a time when computers and 3D models were still unknown.
Accurate knowledge of the location of the underground cavities (e.g. in the form of 3D models) provides an important input for the complex questions that occur in post-mining times. RAG has more than 160,000 2D plans (scanned and georeferenced), which show a total of around 400 seams. The plans contain information on underground mining areas, roadways and shafts.
RAG asks you to develop an MVP that automatically vectorizes the central axes of the mining routes, so that a 3D model can subsequently be created by RAG. 3D models help determine complex interrelationships in order to effectively address arising questions. Younger employees and career changers will especially benefit from 3D models as those facilitate the transfer of know-how.
You will start with high-quality images, so that your image processing algorithm can be trained with relative ease. The Z-component of the route (i.e. its geographical height) can be determined from the height labelling on the 2D plans. The layout of the plans is identical since they are generated by computer software. In this first step, you will also receive an existing 3D model to verify your spatial vectorization. The aim is to replicate the existing model with an accuracy of five meters. Once the algorithm and MVP are fine-tuned, you will try to vectorize plans of poorer image quality. The procedure remains the same, but the pattern recognition algorithm will have to be refined.