Design Of Coal And Gangue Sorting System Based On Ground Penetrating Radar Image

In the process of coal mining and excavation,some impurities such as vermiculite are often mixed,and the calcination value of vermiculite is low.When mixed with coal,it will affect the calorific value and quality of coal.Coal preparation before coal burning has important practical significance for energy conservation.At present,the mainstream coal preparation methods are mostly heavy medium method and jigging method.These two methods have the disadvantages of wasting water resources,polluting the environment,and requiring large plants.In this thesis,based on a large number of relevant data and a large number of experiments to collect radar images corresponding to coal and vermiculite and detailed analysis,the method of identifying coal and vermiculite based on ground penetrating radar image is studied,and the corresponding sorting system is designed.Better practical significance.Due to the different dielectric constants of coal and vermiculite,when the electromagnetic wave emitted by the ground penetrating radar passes through the interface between the medium and the air,the intensity of the reflected electromagnetic waves is different,and the corresponding ground penetrating radar image will be different.In this thesis,the radar images of coal and vermiculite are collected.The feature points method,contrast method and machine learning method are used to analyze the image difference,and the influence of humidity on coal and vermiculite images is proposed.The coal and meteorite recognition algorithm based on logistic regression is proposed.The area of the radar image that can represent the coal and meteorite information is roughly estimated by the height of the coal block and the meteorite and the ground penetrating radar.The exact boundary is determined by the binarized original image,and the area is segmented to extract the feature value for training.The algorithm has the characteristics of being suitable for two classifications,requiring a small number of samples,and a relatively fast speed.Changes in the humidity of coal and meteorites cause changes in radar images.Multiple sets of samples can be trained for different environmental humidity to solve the impact of environmental humidity changes on the recognition results.A coal gangue sorting system is designed,including a queuing module,an acquisition module,an identification module,and a sorting module.The queuing module first screens out the coal and meteorites that are sized to be sent by the queuing machine and conveyor belt to the ground penetrating radar for scanning.The acquisition module obtains the original radar image of the coal block and the meteorite through the acquisition software,and then divides the image area capable of expressing the information.The recognition module trains the sample through a logistic regression algorithm,and uses the training result as the basis for image recognition.The sorting module performs sorting according to the image recognition result,and the baffle at the end of the conveyor belt is controlled by the PLC.Analysis and test results show that radar images can be applied to the separation of coal and vermiculite.