• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Construction research congress >Vision-based recognition of dirt loading cycles in construction sites
【24h】

Vision-based recognition of dirt loading cycles in construction sites

机译：基于视觉的建筑工地污垢装载循环识别

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Automated control of production lines in different segments of manufacturing has been advanced due to emergence of sensing devices and the repetitive character of the processes. The construction industry, however, still suffers from inefficiencies in real-time control of activities due to the manual practice of monitoring, and the fragmented and temporary nature of construction projects. Several sensing technologies including computer vision-based systems have been introduced to address this issue on construction sites. Earthmoving projects are one of the most suitable areas to employ vision-based techniques to extract productivity data because it is possible to select clear sightlines and earthmoving equipment are relatively easy to recognize. In addition to challenges of developing efficient object detection and tracking algorithms, activity recognition based on collected data from detection and tracking engines is yet to be tackled. In this paper, a logical framework is introduced that combines object recognition, tracking, and rational events to recognize dirt loading to a dump truck by a hydraulic excavators, and measure the working cycles and idle times of the earthmoving plants. This logical algorithm showed promising performance in which the variations were caused by deficiencies of recognition and tracking engines rather than the activity recognition algorithm.
机译：由于感测装置的出现和过程的重复特征，已经前进了制造业的不同段中生产线的自动控制。然而，建筑业仍然受到对监测的手动实践的实时控制的效率低下，以及建筑项目的分散和临时性质。已经引入了几种包括基于计算机视觉系统的传感技术，以解决施工地点的这个问题。地球化项目是采用基于视觉的技术的最合适的领域之一，以提取生产力数据，因为可以选择清晰的视野和地球移动设备相对容易识别。除了开发有效的对象检测和跟踪算法的挑战之外，还尚未解决基于来自检测和跟踪引擎的收集数据的活动识别。在本文中，介绍了一种逻辑框架，其结合了物体识别，跟踪和理性事件，以通过液压挖掘机识别到自卸车的污垢装载，并测量地耕植物的工作周期和空闲时间。这种逻辑算法显示出具有有希望的性能，其中变化是由识别和跟踪引擎的缺陷而不是活动识别算法引起的。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号