Real-time Monitoring And Early Warning Roadway Deformation Based On Machine Vision

The deformation and failure of roadway can disturb the underground traffic,damage production equipments, cause casualties and threat seriously to safety andproduction in coalmines. Meanwhile, the inoculation and occurrence of coal or rockdynamical disasters associates with the deformation and failure of roadway. Roadwaydeformation monitoring can make early warning for surrounding rock instability,reveal space-time evolution rules of roadway deformation, and provide evidence forroadway support, prevention of coal or rock dynamical disasters and mining pressureand strata movement rules.According to the monitoring and early warning of roadway deformation, thispaper studied the real-time monitoring experimental system of rock deformation,studied the evolution law of deformation indexes of coal, rock and concrete underingloading, and studied the dynamic characteristics of deformation and failure of rockaffected by mining, established the multi-index early warning rule of roadwaydeformation and failure and the field application system based on the machine visionprinciple. The main results of this paper described as follows:(1)This paper experimental studied the deformation and failure law of coal,rock or concrete under uniaxial loading, the results show that the samples present theregular dynamic characteristics before failure; constructed the coal, rock or concretedamage model including strain and damage degree based on the damage statisticalmechanics method; and established the relation between macroscopic deformation andmeso-damage; described the process of sample deformation and failure bydeformation size, deformation velocity and deformation acceleration and definedevery critical value in each stage.(2)Based on the machine vision monitoring principle, this paper developed thereal-time deformation monitoring experimental system for coal and rock mass anddeveloped the corresponding software, which can achieve the full field of view,continuous, non-contact and on-line monitoring for two-dimensional andthree-dimensional deformation or movement, and meet the needs of experimentalstudy for dynamic characteristics of surrounding rock deformation. Based on real-timedeformation monitoring experimental system, this paper studied the surrounding rockdeformation process after excavating and different surrounding rock deformationunder different loads by similarity simulation experiment method; studied the roof deformation and caving process during the coal seam mining, and obtained theevolution law of monitoring index of surrounding rock deformation, and establishedthe corresponding relation between dynamic characteristics of surrounding rockdeformation and surrounding rock instability.(3)According to the problems such as more missing report and poorapplicability in existing surrounding rock stability criterion, this paper established thecriterion for comprehensive evaluation of surrounding rock stability by deformationsize, deformation velocity and deformation acceleration; proposed the real-time,regional roadway deformation monitoring method based on the machine visionprinciple aiming at the continuous and regional character of roadway deformation,which can automatically collect and process the index data of roadway deformationand judge the surrounding rock stability by dynamic characteristics of surroundingrock deformation.(4)This paper established the roadway deformation monitoring system used infiled based on the machine vision monitoring principle, tested the deformation processof gate roadway, and had early warning to unstable failure of surrounding rock, theresults show that the deformation indexes reflected the state and trend of surroundingrock, and the dynamic characteristics of surrounding rock deformation wellcorresponded to the stress concentration in time distribution, and the deformationindexes can be the precursor information for surrounding rock instability.