摘要
热电厂存煤量的实时检测是数字化生产的重要环节,检测得到的大量统计数据为煤量供耗科学管理提供数据支撑。提出一种图像和激光相结合的低成本运煤车厢物料监测和非接触式体积测量方法。首先,基于神经网络的目标检测方法和K-Means聚类算法实现对进出煤场车厢类型和标号的自动识别;其次,利用光流法获取车厢运动速度信息,采用二维激光和图像测速相结合的方式生成车辆的三维点云模型;最后,利用区域生长方法,得到特定区域内的有效激光点云数据,并利用多重积分和三维重构技术,得到不同车厢的装煤量体积数据。现场试验结果表明,该方法可以对不同运煤火车车厢装煤量进行实时测量,测量精度达到92%以上。
The real-time detection of coal storage in thermal power plant is an important part of digital production.A large number of statistical data obtained from the detection provide data support for the scientific management of coal supply and consumption.A low-cost material monitoring and non-contact volume measurement method for coal carriage based on image and laser was proposed.Firstly,the target detection method based on neural network and K-means was proposed The clustering algorithm realized the automatic recognition of the type and label of the cars in and out of the coal yard.Secondly,the optical flow method was used to obtain the moving speed information of the carriage,and the 3D point cloud model of the vehicle was generated by combining the two-dimensional laser and image speed measurement.Finally,the effective laser point cloud data in a specific area was obtained by using the region growing method,and the multiple integration and three-dimensional reconstruction technology were used,The coal loading volume data of different cars were obtained.The experimental results showed that this method could be used to measure the coal loading of different railway cars in real time,and the measurement accuracy was more than 92%.
作者
卢进南
韩建国
王常宝
郭友瑞
王志良
LU Jinnan;HAN Jianguo;WANG Changbao;GUO Yourui;WANG Zhiliang(College of Mechanical Engineering, Liaoning Technical University, Fuxin 123000, China;Kailuan (Group) Co., Ltd., Tangshan 063018, China;Tangshan Kailuan Thermal Power Co., Ltd., Tangshan 063018, China)
出处
《煤矿机电》
2021年第1期49-53,56,共6页
Colliery Mechanical & Electrical Technology
关键词
存煤量
神经网络
激光点云
体积检测
coal storage
neural network
laser point cloud
volume measurement
