Font Size: a A A

Research On Recognition And Dynamic Characteristics Of Coal Wall Spalling Based On Machine Vision

Posted on:2022-04-21Degree:MasterType:Thesis
Country:ChinaCandidate:Y H LiFull Text:PDF
GTID:2481306533469374Subject:Mining engineering
Abstract/Summary:PDF Full Text Request
Coal wall spalling is one of the common underground disasters in coal mine.It threatens the safety of coal miners,damages the hydraulic support,shearer,scraper conveyor and other mining equipment,and threatens the production safety of coal mine.Based on the monitoring image of 8222 working face and the resistance data of hydraulic support in Tashan mine,this paper uses theoretical analysis,simulation experiment,data analysis and other methods to de blur the monitoring image of working face,and studies the motion detection and dynamic characteristics of rib spalling.The main contents and conclusions are as follows:(1)The characteristics of coal wall image of underground working face are analyzed and studied.It is found that coal wall image is mainly affected by coal dust and uneven illumination.For adverse factors,different applicable algorithms are analyzed.Finally,the coal wall image deblurring model based on generative confrontation neural network is determined and built.Through comparison with other algorithms,the algorithm in this paper is better.(2)According to the dynamic characteristics of coal wall spalling,three kinds of moving target detection algorithms are compared and analyzed,and a coal wall spalling motion detection algorithm based on SSIM and background difference method is designed,and the video data of spalling recognition experiment is processed.The experimental results show that the algorithm can accurately record the whole process of spalling and collect the time,area and change curve of spalling Information.(3)Through the analysis and Research on the dynamic characteristics of coal wall spalling,it is found that the stability of coal wall in the middle of the working face is poor,and it is more prone to occur coal wall spalling;before the occurrence of spalling,there will be precursor information such as block sliding and coal wall fracture,and the time interval from the occurrence of precursor to the occurrence of spalling is from 3 minutes to 2 hours.(4)It is found that spalling mainly occurs in the period of roof weighting,the probability is 58.8%,the probability before roof weighting is 11.8%,the probability after roof weighting is 17.6%,and the probability of pressure peak area in non roof weighting stage is 11.8%.In the pressure rise area,the support resistance is not less than 16.2mpa when the rib spalling occurs;in the pressure drop area,the minimum support resistance is 16.4mpa when the rib spalling occurs.There are 57 figures,11 tables and 85 references in this paper.
Keywords/Search Tags:coal wall spalling, machine vision, neural network, object detection
PDF Full Text Request
Related items