| Large geographical area of coal mine,numerous winding and crossing underground roadways,complex and rugged terrain caused by serious destruction of roadways after the disaster,uneven distribution of explosive gas and dust in the air after the disaster,dark and humid visual environment,underground environment greatly limited wireless communication,these environmental factors caused great difficulties to the walking and navigation of robots.After years of efforts,China University of Mining and Technology has obtained a lot of achievements in the explosion-proof design and walking mechanism of coal mine rescue robot,and obtained the certification of coal mine safety through the application demonstration of major projects,solved the problem of remote control walking of coal mine rescue robot basically.But the in-depth research is needed in underground navigation.The paper is carried out for improving the visual navigation ability research of the coal mine rescue robot by visualization technology.Firstly,this paper carries out research on the environmental structure characteristics of underground coal mines,including structure and topographic characteristics of roadways before and after disasters,visual environment characteristics of roadways and electromagnetic environment characteristics of roadways.According to the characteristics of the overall structure of the coal mine,this paper divides the whole rescue process of the coal mine rescue robot into five stages for analysis,and summarizes the environmental limitations and corresponding countermeasures faced by the underground navigation in different stages.According to the characteristics of underground visual environment such as zero illumination,dust and humidity,the influence of zero illumination on visual sensing is analyzed.According to the structure of underground mine roadway and the special electromagnetic environment caused by various equipment,the influence of electromagnetic interference and wireless communication characteristics in roadway on the robot is analyzed.Then,aiming at the problem of image degradation caused by the influence of visual environment factors in mine disaster relief robot's return video,this paper proposes to use image enhancement to improve the visual ability of the robot.The gray histogram stretching and equalization algorithm are used to solve the problem of zero illumination,the dark channel defogging algorithm is used to solve problems of the dust and smoke interference,the image stabilization algorithm is used to solve the problem of image jitter caused by the bump in the robot walking,and the thermal imager is used to detect the life and heating equipment,and the depth camera is used to detect the topographic structure under logging.Then,aiming at the problem of position visualization of robot walking in large area of underground coal mine,the virtual reality technology is used to realize the three-dimensional visualization display of the whole coal mine,and the real-time walking position and space trajectory of the robot are identified.In addition to the overall three-dimensional model of the coal mine,modular stacking technology is also uses to realize the rapid construction of the mine underground disaster area roadway model in this paper,combined with embedded RFID tags and reflective sign recognition methods to solve the underground geographic area positioning problem.The improved A* algorithm is used to realize path navigation in virtual three-dimensional space,and the robot's three-dimensional visual navigation is realized by combining the robot's video remote control navigation.Then,this paper realizes the visual navigation test platform which combines virtual and real,and solves the problem that the test prototype can not go down well to carry out a lot of test work because of problems of the safety and regulations.The whole platform is divided into three main parts,mobile test platform of visual navigation algorithm,virtual visual navigation test system and optical motion position capture system platform.The platform combines physical robots and sensors,virtual robots and sensors and carries out joint debugging.Finally,three comprehensive experiments are carried out in this paper in order to validate the virtual experiment platform and video enhancement technology in visual navigation research,which are underground visual environment virtual experiment,visual navigation experiment in application demonstration and virtual reality-assisted visual navigation experiment in roadway.In summary,this paper provides some new ideas for the underground navigation of coal mine disaster rescue robot through the research of some key technologies of underground visual navigation of coal mine disaster relief robot. |
PDF Full Text Request