| Coal industry is basic industry of China,and coal mine safety has always been a key issue for researchers.When it comes to inspection by workman,there are several difficulties such as labor intensity,high rate of missed inspection,poor safety and so on.The use of robot for intelligent inspection and monitoring of coal mine improves the automation and intelligent level of coal mine,reduces the workload of inspection personnel as well as improves the efficiency of underground safety inspection,and it has become one of the mainly researched directions of coal mine safety.The available inspection scheme of underground roadway of coal mine at present is mainly based on manual remote monitoring,which is easily affected by underground smoke and light,and it has low efficiency.In this thesis,a solution is proposed for the video inspection of structural abnormalities in underground roadway of coal mine,and then a coal mine roadway inspection robot system is designed.The main contents of this thesis are as follows:(1)The platform of underground inspection robot is designed and established.The orbital inspection robot for underground roadway of coal mine is selected through the investigations into the scene and requirements of coal mine inspection robot,the overall structure of the robot is designed based on this,and then the hardware system of the robot is constructed.The central control module of the robot is designed based on STM32F407IGT6,and the robot contains power management module,motion control module,information acquisition module and wireless communication module and so on.The control software is designed based on the human-computer interaction function,and the whole coal mine roadway inspection robot system is developed.(2)The software and hardware system of inspection robot for underground roadway of coal mine is designed.The cruise logic of the inspection robot is designed according to the requirements of the robot.On one hand,the cruise scheme including ultrasonic obstacle avoidance and robot positioning function is designed;on the other hand,the inspection video acquisition scheme combining with Radio Frequency Positioning System is designed which aimed at the inspection video anomaly detection algorithm.The software system of the robot is designed based on inspection and users requirements,the embedded control software supporting daily inspection and the host monitoring software are included.(3)A video anomaly detection algorithm for underground inspection is proposed.Because the underground intelligent video monitoring technology is still limited to the intelligent processing of fixed monitoring so far,and it is difficult to meet the needs of intelligent inspection of underground roadway of coal mine.An improved background difference detection algorithm for underground inspection video is proposed based on the idea of fixed monitoring background subtraction algorithm.According to the characteristics of the inspection video collected by the track inspection robot,the method of establishing the background model is designed based on the keyframes.Taking the keyframes as the background reference,mean hash was used to find the background model of the keyframes in the inspection video to be detected,so that the sampling keyframes can be used as the input to detect the inspection video by background subtraction.The anomaly targets are marked and the then background model is updated eventually according to the detection results.(4)The robot function module and inspection video anomaly detection algorithm are tested and analyzed through the inspection robot prototype.The experimental results show that the inspection robot could meet the requirements of preset underground inspection tasks,and the anomaly detection algorithm of inspection video shows good performance. |
PDF Full Text Request