The construction of Norwegian railway network started in 1851, and now it has 4200km of rail, 2700 bridges. The need to improve condition assessment and predictive maintenance of railway infrastructure is essential to a sustainable society when facing the fast deteriorating and aging infrastructure and increased demands from environmental and operational loading. Currently, this is primarily achieved by yearly visual inspection by human experts. This is a tremendous task given the number of bridges and components in these bridges and have huge economic cost associated with it. During the last five years, large advances in the field of computer vision and machine learning has been made and become cheaper and widely available through consumer-grade technology. Adopting these technologies has the potential to reduce cost associated with maintaining the infrastructure and improving the inspections for better assessment and use of existing structures.
The project is focused on an advanced monitoring method, a computer vision-based inspection system. We will use deep learning-based method to identify different structural components and to evaluate structural damage from images or videos.