Intelligent identification of structural planes of tunnel surrounding rock based on three-dimensional point cloud

Structural planes in surrounding rock of tunnels are one of the main factors affecting the stability of surrounding rock. Traditional methods to obtain these structural planes have shortcomings such as low work efficiency, large workload, and strong subjectivity. To address issues like reconstruction of real-scene tunnel surrounding rock models, 3D point cloud extraction, and structural plane recognition, a self-supplementary lighting digital image acquisition device for underground tunnel chamber surrounding rock was developed to improve the quality of tunnel surrounding rock image acquisition. Using the SFM algorithm, a 3D real-scene model of the tunnel surrounding rock was reconstructed, and point cloud data were extracted. An algorithm was also proposed to identify rock mass structural planes based on the normal vector features of the tunnel surrounding rock point cloud. Finally a semi-automated identification system for tunnel surrounding rock structural surfaces was developed. Taking a specific tunnel as an example, we verified the applicability of the invented apparatus, and its accuracy can meet engineering requirements. However, misidentifications can not be ignored when identifying structural planes with a few quantities or manifested by traces, faulty identification is inevitable, which could be solved by further algorithm optimization.