视觉大模型在轨交隧道缺陷检测中的应用研究
Application of Large Vision Model for Defect Detection in Orbital Tunnel
方爱国1
作者信息
- 1. 上海点泽智能科技有限公司,上海 200333
- 折叠
摘要
本文针对轨交隧道巡检过程中,人工巡检工作存在的效率低、容易遗漏等问题,提出采用轨交检测车采集隧道内壁信息,结合轨交隧道的相关巡检规范,基于V-MoE视觉大模型开发轨交视觉大模型.与传统的机器视觉相比,在隧道壁潜在风险和缺陷检测方面,该方法的识别类型和识别精度均大幅提升,可以有效提升识别准确率,是视觉大模型在新领域的重要突破.
Abstract
This article addresses the issues of low efficiency and easy omission in manual inspection during the inspection process of rail transit tunnels.It proposes a rail transit visual model developed based on the V-MoE visual model,which uses rail transit inspection vehicles to collect information on the inner walls of tunnels and combines with relevant inspection standards for rail transit tunnels.Compared with traditional machine vision,this method significantly improves the recognition type and accuracy in detecting potential risks and defects on tunnel walls,effectively improving recognition accuracy.It is an important breakthrough in the new field of visual modeling.
关键词
视觉大模型/V-MoE/轨道交通/隧道壁缺陷检测Key words
large vision model/V-MoE/rail transit/tunnel wall defect detection引用本文复制引用
出版年
2024
NETL
NSTL
万方数据