| Coal mine is the country’s important foundation industry,with the development of mining technologies,although the safety situation of coal mine is gradually getting better,but compared with other industries,coal mine safety accident is still the major problem faced by our country coal mine safety management.Facing the increasingly severe situation of coal mine safety production,how to discover the hidden danger of accidents in advance and improve the efficiency of coal mining is an urgent problem to be solved at present.Mine belt conveyor is an important large mechanical and electrical equipment of coal transportation,the main artery of coal transportation,and also the frequent occurrence of coal mine safety accidents.With the rapid advancement of industrial automation and information technology,the daily production of coal needs to grow,which for the belt conveyor coal load and transport volume also increased.However,in the process of coal transportation by the belt conveyor,coal flow often contains coal gangue,large coal,bolt and other foreign bodies,and in the process of coal transportation,the surrounding area of the belt conveyor also has security risks such as personnel work violations,resulting in belt tearing,coal stacking,coal blocking,casualties and other accidents.Therefore,this paper focuses on the safety monitoring and protection of belt conveyor,carries out intelligent video analysis on the surveillance video of belt conveyor through the design of target detection and identification network model,and establishes the safety monitoring and protection system of belt conveyor from the source.The main research contents and innovations of this paper are as follows:(1)Proposed a coal flow foreign body detection method LE-YOLOv3 under the condition of uneven illumination of mine artificial lighting,and the system development and field test application are also carried out.Firstly,the actual lighting environment of underground coal mine is analyzed.Then,based on the target detection algorithm model YOLOv3,which is suitable for non-coal foreign matter detection,an improved YOLOv3 foreign matter detection algorithm LE-YOLOv3 oriented to the enhancement of non-uniform illumination edge of mine is proposed to improve the accuracy of coal flow inclusion foreign matter detection of mine belt conveyor.Firstly,the EEM edge enhancement module was added to the LLNet structure,and then the fusion module was further combined with the network framework of YOLOv3 to enhance the illumination information of coal flow monitoring image of mine belt conveyor.Then,the bolt data set samples were enhanced by the mixed data set enhancement method of linear difference and random clipping,which significantly improved the foreign body detection effect.Finally,through the interface design and system development,and the test and analysis on the coal mine production site,it is proved that the proposed detection method can effectively detect and identify foreign bodies in coal flow.(2)PS-YOLOv3,a target detection algorithm oriented to the complex environment of mine main transportation roadway,is proposed and applied to realtime monitoring of illegal behaviors such as personnel wearing safety helmets in the operation working area of mine belt conveyor and intrusion in dangerous areas.Firstly,the PS multi-level feature fusion module is added to the original YOLOv3 algorithm model,and the different levels of features extracted from the backbone network are aggregated.At the same time,SE attention module is introduced into the Neck layer in PS-YOLOv3,so as to pay attention to the salient characteristics of targets at different scales,improve the detection ability of the model.Finally,through the client interface design and system development,it is also applied to the coal mine production site.The results show that the proposed method and system in this paper can achieve remarkable detection effect and practical value.This paper has 32 figures,3 tables and 85 references. |
PDF Full Text Request