Abstract: Objective Current comprehensive 3D tunnel structure inspection features low efficiency and speed, unable to meet the operational and maintenance needs of high-density rail transit networks. Moreover, the limited intelligent management capabilities for heterogeneous data such as point clouds and images restrict the in-depth exploitation of inspection data. Therefore, it is necessary to conduct research on tunnel 3D structure high-speed inspection systems and key data processing technologies. Method Using high-speed laser scanners based on modular laser ranging unit array combined with inertial navigation units, a tunnel 3D structure high-speed inspection system is developed. A progressive ellipse-fitting algorithm is developed to iteratively filter noise and fit tunnel cross-sections for tunnel localized deformation calculation. A multi-ring point cloud segmentation fitting and optimal value selection strategy is adopted to reduce the support and pipeline interference and improve the accuracy of segment deformation analysis. A tunnel inspection data platform is built to enable integrated display and management of point cloud and image data, and to support tunnel health assessment based on operation-maintenance demands. Result ＆ Conclusion Through practical engineering applications, it is verified that the tunnel 3D structure high-speed inspection system can comprehensively acquire tunnel point cloud and image data at speeds of no less than 80 km/h, and can intelligently detect structural and surface defects, thereby significantly enhances the automation and intelligence levels of comprehensive tunnel inspections.