BUSINESS SCOPE
3D Fine-Grained Integrated Advanced Geological Forecasting and Intelligent Identification Technology for Highway Tunnels and Geological Bodies
Project Overview:
This project leverages image recognition technology based on artificial intelligence algorithms and multivariate data-fitting methods from machine learning. It integrates geological sketches, geophysical exploration results, borehole parameters, and monitoring measurement data obtained from advanced geological forecasting and monitoring during tunnel construction to establish an intelligent image recognition and multivariate data-fitting model for the tunnel face. The model employs both one-stage object detection algorithms (including YOLO series, SSD series, and RetinaNet) and two-stage object detection algorithms (including R-CNN, Fast R-CNN, Faster R-CNN, and Mask R-CNN). Specifically, it adopts the Mask R-CNN instance segmentation algorithm from the two-stage approach. Combining the fundamental theories of tunnel surrounding rock classification with the distinctive features of AI algorithms, this project proposes an intelligent method for classifying tunnel surrounding rocks by integrating multi-source heterogeneous data, including geological surveys, advanced geological forecasting, borehole drilling data, and geological sketches of the tunnel face. A comprehensive framework has been designed and implemented in a computer program, enabling the integrated determination of the surrounding rock classification at the tunnel face.
Project Highlights:
1. Based on the optimized combination of artificial intelligence algorithms and various geophysical exploration methods, this project proposes an intelligent identification and classification method for surrounding rock at tunnel face based on deep learning, an intelligent recognition technique for unfavorable geological conditions in tunnel radar images, and a comprehensive, three-dimensional, high-resolution advanced geological forecasting technology that integrates both in-tunnel and out-of-tunnel data. It establishes a comprehensive forecasting system and model tailored to different types of unfavorable geological conditions, and has developed an advanced geological forecasting system for highway tunnels. This system provides strong technical support for reducing the likelihood of geological hazards during tunnel construction, ensuring tunnel construction safety, and enabling “dynamic design adjustments, dynamic construction, and reasonable savings in schedule and investment.”
2. An intelligent image recognition and multi-dimensional data fitting model for tunnel face images has been established. A method for classifying tunnel surrounding rock based on multi-source heterogeneous data has been proposed, enabling integrated intelligent identification and comprehensive classification of the surrounding rock at the tunnel face.
3. For high-risk sections in tunnel construction, a three-dimensional, refined, and integrated geological advance forecasting model combining both in-tunnel and out-of-tunnel approaches has been proposed. Based on the physical property characteristics of different types of unfavorable geological conditions (such as karst, fault fracture zones, and water-rich areas) as well as the strengths and weaknesses and parameter suitability of various geophysical exploration methods, a comprehensive forecasting system tailored to specific types of unfavorable geological bodies—including karst, fault fracture zones, and water-rich areas—has been established, thereby enhancing the accuracy of geological forecasting.
4. We have developed an advanced geological forecasting system for highway tunnels, enabling timely uploading of geological forecast information, prompt feedback, rapid early warning, and swift response. This has effectively enhanced the level of information-based management and overall effectiveness of tunnel geological forecasting projects.
Honored Achievements:
Key Promoted Achievements of the “Four New Technologies” for Highway Tunnels by the China Highway Society in 2025.
Address: No. 100, Yangguan Avenue, National High-tech Industrial Development Zone, Guiyang City, Guizhou Province
mobile client
