基于机器视觉的煤炭排矸系统设计

Design of Coal Gangue Discharge System Based on Machine Vision

针对传统煤矸石分拣采用人工方式,错选率高、效率低等问题,设计了基于机器视觉的煤炭排矸系统。利用YOLOv3+VGG16卷积神经网络对煤矸石进行两级目标检测,采用三维执行机构实现煤矸石自动抓取,设计控制方法协调系统各模块工作。装置加装在带式输机上进行试验,煤矸石分选的正确率达到96%以上,对比人工方式效率有明显的提升,证明系统具有较强的应用潜力。

Aiming at the problems of high error rate and low efficiency in traditional coal gangue sorting,a coal gangue discharging system based on machine vision is designed.Use YOLOv3+VGG16 convolutional neural network to detect the coal gangue at two levels,use three-dimensional actuator to realize automatic grab of coal gangue,and design control methods to coordinate the work of each module of the system.The device is installed on the belt conveyor for test,and the correct rate of coal gangue separation is more than 96%.Compared with the manual method,the efficiency is significantly improved,which proves that the system has strong application potential.