Tunnel inrush water measuring method based on multi-view fusion and optical flow field calculation

By using multi-view fusion and optical flow field calculation, deep learning and optical flow models are used to automatically extract the water inrush area in the tunnel, which solves the problems of low efficiency and insufficient accuracy of traditional tunnel water inrush monitoring and realizes high-precision automated measurement.

CN121999232BActive Publication Date: 2026-06-26CHENGDU UNIVERSITY OF TECHNOLOGY
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
CHENGDU UNIVERSITY OF TECHNOLOGY
Filing Date
2026-04-10
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

Traditional tunnel water inflow monitoring relies on manual observation, which is inefficient, has limited accuracy, and poses safety hazards. Existing non-contact methods are not robust to low-texture complex backgrounds due to insufficient single-view information, making it difficult to achieve high-precision flow rate measurement.

Method used

The method employs multi-view fusion and optical flow field calculation. It automatically extracts the inrush water region through a deep learning segmentation model and combines it with an advanced optical flow field calculation model to calculate the flow rate. This includes multi-view video shooting, image preprocessing, U-Net semantic segmentation, and RAFT optical flow model calculation. The multi-view velocity vector field is fused to calculate the inrush water volume.

Benefits of technology

It achieves non-contact, high-precision, and automated calculation of tunnel water inflow, overcoming the effects of insufficient lighting and complex working conditions, and improving the accuracy and stability of the measurement.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121999232B_ABST
    Figure CN121999232B_ABST
Patent Text Reader

Abstract

The application relates to a tunnel gushing water measuring method based on multi-view fusion and optical flow field calculation, relates to the technical field of image recognition, shoots a tunnel face gushing water video, carries out image frame extraction on the obtained video, carries out pretreatment on the frame-extracted images, establishes a tunnel face image data set, labels and data enhances the pretreated images, trains a U-Net semantic segmentation model to identify gushing water area images in the image data set, adopts a RAFT optical flow model to calculate adjacent frame pixel displacement fields in the gushing water area images in the images, extracts multi-view effective flow velocity vector fields, fuses the multi-view effective flow velocity vector fields, and calculates the tunnel face gushing water volume; the application adopts U-Net to carry out pixel-level segmentation on the gushing water area through multi-view imaging, adopts the RAFT optical flow with a mask constraint to carry out velocity measurement, carries out velocity vector fusion on a unified reference plane, can overcome the influence of insufficient light, reflection and local shielding in the tunnel, and improves the accuracy and stability of the result.
Need to check novelty before this filing date? Find Prior Art