ABSTRACT
Faulting along the HS2 Chilterns tunnel alignment is underreported, so remote sensing data is used to identify the location of faults. The growing accessibility of remote sensing data means it is increasingly being harnessed for site investigation. However, this data is often large and time intensive to interrogate manually. As such, computer vision techniques can be employed to automate the identification of faults and tectonic lineaments along linear infrastructure in LiDAR-generated digital terrain models (DTM). This work aims to investigate how the surface expression of a fault can be seen in a region’s geomorphology using LiDAR data and to generate a DTM creating a script that automates fault identification using computer vision and evaluates the performance of the workflow against manually identified faulting. The algorithm is susceptible to man-made features and performs better at larger scales. Most manually identified faults had corresponding tectonic lineaments; however, not every detected line segment corresponded to faulting. The results suggest that automated lineament extraction can be an effective tool for preliminary site investigation but should not be used without subsequent ground investigation to confirm any conclusions drawn.
