Coal Flow Monitoring System Of Belt Conveyor With Integrated Perception

As a basic energy source,coal occupies an important position in China's economic development.With the recovery of the domestic coal industry in the past two years,the state has increasingly paid attention to coal mine safety,and supervision has been continuously strengthened.Large and medium-sized coal mines already have coal mine safety monitoring.Unattended intelligent coal mines have developed rapidly.Belt conveyors play an important role in the transportation of coal mines.However,belt conveyors usually run at a constant speed and it is difficult to achieve full load.Therefore,so the phenomenon of "big horse pulling small car" often occurs.Some non-coal foreign objects often appear on the belts of belt conveyors,such as:sleepers,anchors,large pieces of vermiculite,etc.These foreign bodies can easily cause the belt to break and even cause coal mine accidents,which not only affects the normal production of the coal mine,but also causes serious economic losses.This thesis studies the coal flow monitoring system of belt conveyors based on information fusion.Based on real-time video monitoring of coal mines,it mainly uses deep learning and intelligent control technologies to propose lightweight target detection on the embedded platform NVIDIA Jetson Nano,and speed regulation of belt conveyor.Aiming at the problem of low-quality monitoring images of coal mine underground belt conveyors,based on the traditional Retinex image enhancement algorithm,combined with the idea of CNN(Convolutional Neural Network),an adaptive underground environment image enhancement algorithm RS-Net is proposed.In order to apply the coal flow detection neural network to the underground environment of the coal mine,this thesis proposes a lightweight deep learning target detection network Dense-YOLO,which can quickly identify coal flow and non-coal foreign matter on an embedded platform,and based on The coal flow rate is calculated to match the speed of the belt conveyor.In order to improve the real-time performance of the belt conveyor speed control system,it is proposed to use its controller to build a belt conveyor speed control system on the same embedded platform.The speed control system first interweaves coal flow and belt speed information to the embedded platform control Controller,then the controller sends a speed adjustment signal to the inverter,and finally the inverter completes the speed regulation of the motor.Finally,the thesis tests the key technologies of the system and analyzes the test results.Experimental results show that,the Dense-YOLO+Retinex target detection method proposed in this thesis,whether on a computer or on an embedded platform,it can effectively improve the target detection recognition rate in the monitoring image of a coal mine underground belt conveyor.It has high application value in coal flow identification and non-coal foreign matter identification of belt conveyors in coal mines.It has laid a good foundation for achieving energy saving and consumption reduction in coal mines and safe production in coal mines,and has good promotion value.Figure[36]table[12]reference[86]...