煤炭工程 ›› 2024, Vol. 56 ›› Issue (9): 202-210.doi: 10.11799/ce202409030

• 研究探讨 • 上一篇    下一篇

基于TOF 相机的振动筛上物料状态动态监测方法

卢 军,谭兴富,成举炳,华智诚,刘 畅   

  1. 1. 中煤科工集团北京华宇工程有限公司,河南 平顶山 467000

    2. 中国矿业大学(北京)化学与环境工程学院,北京 100083

  • 收稿日期:2023-11-08 修回日期:2024-06-03 出版日期:2023-09-20 发布日期:2025-01-08
  • 通讯作者: 卢军 E-mail:hzcheng0812@163.com

Dynamic monitoring method for material status on vibrating screen surface via TOF camera

  • Received:2023-11-08 Revised:2024-06-03 Online:2023-09-20 Published:2025-01-08

摘要:

振动筛筛上物料载荷异常不仅影响生产效率还可能导致安全隐患,实时准确地监测筛上物料状态具有重要意义。针对传统人工监测方法劳动强度大、速度慢、有极高安全隐患并且不容易发现设备微小故障等问题。基于TOF相机物料动态监测技术,提出了一种联合KD树与帧差法的点云数据补全方法和基于曲面重构的筛上物料动态监测方法。该点云数据补全方法有效修补了因筛上物料表面水分和物料缝隙导致的TOF相机拍摄过程中的点云数据缺失问题,通过对补全后的筛上物料进行点云曲面重构分析,判断欠载、过载、左偏载和右偏载四种筛上物料状态,实时监测筛上物料状态。研究通过TOF相机直接获取筛上物料状态信息,对筛上物料实时检测,减少了人工干预的需求且具备实时性。为筛上物料状态实时监测问题提供了一个可靠的解决方案。

关键词: 振动筛, TOF相机, 3D点云, 计算机视觉, 筛面监测

Abstract:

Abnormal material loading on the vibrating screen is a common issue in industrial production, with repercussions that extend beyond production efficiency to include potential safety hazards. Therefore, monitoring material status on the screen in real-time and with precision holds immense importance. However, the traditional monitoring method primarily relies on manual on-site observation, which is labor-intensive, slow, fraught with significant safety risks, and often inadequate for identifying minor equipment faults. This approach falls short of the demands for material status monitoring on the screen. This paper introduces a dynamic material status monitoring method for vibrating screens that utilizes a Time-of-Flight (TOF) camera. This method combines a point cloud data completion approach with a KD tree within the frame difference method, alongside a dynamic monitoring technique based on surface reconstruction. These methodologies enable the real-time monitoring of material status on the screen. The point cloud data completion method, coupled with the KD tree within the frame difference process, effectively addresses the challenge of missing point cloud data caused by moisture on the material's surface and gaps between materials during TOF camera capture. Through the analysis of the reconstructed point cloud surface of the completed screen materials, it becomes possible to determine the status of four material conditions on the screen: underload, overload, left bias load, and right bias load, thus enabling real-time material detection on the screen. This study acquires material state information on the screen directly via the TOF camera, reducing the need for manual intervention and providing real-time performance. Consequently, it offers a highly efficient and dependable solution for real-time monitoring of material conditions on the screen.

中图分类号: