露天矿破碎站链条断裂安全监测技术研究

Study on safety monitoring technology for chain fracture in crushing station of open-pit mine

为解决露天矿中破碎站输送机构链条断裂故障检测效率低,传统图像识别方法处理效率低,检测实时性差的问题,提出一种基于深度学习技术的破碎站链条断裂检测技术。首先,从硬件平台搭建和软件架构上提出链条断裂检测系统总体设计方案;其次,使用YOLOv4模型搭建链条检测模型架构,根据破碎站运行工况提出链条断裂检测算法,并结合实际环境特征提出图像预处理方法;然后,采集图像样本用于模型迭代训练,得到链条目标检测识别模型;最后,检测链条断链丢失情况。结果表明:基于深度学习的破碎站链条目标检测可以在输送机构运行过程中精确识别出图像中链条的数量;当链条被遮挡模拟丢失后,能够及时发现并报警。

The detection efficiency of the chain fracture of the conveying mechanism in the crushing station of the open-pit mine is low,and the processing efficiency of the traditional image recognition method is low,with poor real-time detection performance.To address these issues,a chain fracture detection technology of the crushing station based on deep learning technology was put forward.Firstly,the overall design scheme of the chain fracture detection system was proposed in terms of the hardware platform construction and software architecture.Then,the YOLOv4 model was used to build the chain detection model architecture,and the chain fracture detection algorithm was developed according to the operating conditions of the crushing station.The image preprocessing method was proposed according to the actual environmental characteristics.Then,image samples were collected for iterative training of the model,and the chain target detection and recognition model was obtained.Finally,the chain fracture loss was detected.The results show that the chain target detection of the crushing station based on deep learning can accurately identify the number of chains in the image during the operation of the conveying mechanism,and when the chain is occluded for simulated loss,the model can send a warning in time.