Research & Development

Rail Vision's solutions are heavily underpinned by innovative research. Rail track images are highly variable in content. How they appear depends on type of track, lighting conditions, and camera parameters. Rail Vision over the years has developed a detailed understanding of such challenges and how best to solve them through the optimisation of hardware and software.

Some of the key innovations of such work include the modelling of poor quality image data for classification, extracting three dimensional information from 2D images based on temporal changes and camera lens parameters, and relating detailed semantic knowledge of railways to problem solving in software.