Tunnel face monitoring is of great significance for ensuring construction safety and project quality. However, traditional monitoring methods relying on manual measurement and on-site observation are inefficient and limited in accuracy. To improve the monitoring level, a rapid digitization method of tunnel face combined with deep learning was proposed. By constructing a homography transformation model of the tunnel face, applying a feature matching algorithm based on deep learning and a fast adaptive joint feature extraction algorithm based on multi-scale information, rapid and accurate joint feature extraction independent of camera parameters and shooting angles was achieved. Real-time monitoring was also accomplished with the help of fast joint detection and fusion methods. The experimental results show that this method can effectively extract the global joint information of the tunnel face, significantly improve the data acquisition efficiency and reduce the operation difficulty. In conclusion, this method provides a new idea and approach for tunnel engineering monitoring and effectively guarantees the safety of tunnel construction.