首页|基于机器视觉的船舶管路滴漏监测

基于机器视觉的船舶管路滴漏监测

扫码查看
为在管道泄漏初期及时检测出微小的滴漏故障,提出一种管路滴漏视觉监测模型.采用均值背景差分法检测管路滴漏,提取液滴前景的特征参数,结合虚拟线圈法,基于滴漏液滴个数进行流量统计,提出滴漏体积流量的评估方案.采取搭建滴漏实验台的方法获取滴漏视频对模型进行验证,结果表明,该模型能够很好地检测出管道滴漏液滴,特别是对于滴漏频次较慢的泄漏,能够准确统计出滴漏个数且准确率在98%以上,泄漏体积流量的估计值相对误差在20%以内.该方法可以有效对管路滴漏进行监测,为管路维修决策的制定提供了参考.
Ship pipeline drip monitoring based on machine vision
In order to detect small drip faults in the early stage of pipeline leakage,a visual monitoring model of pipeline drip leakage was proposed.The mean background difference meth-od was used to detect pipeline drip and extract feature param-eters of droplet prospects,and by combining the virtual coil method,an evaluation scheme of drip volume flow rate was proposed based on the number of dripping droplets for flow statistics.Then the drip test bench has been constructed to ob-tain a drip video,and then to validate the model.The results show that the model can effectively detect pipeline drip drop-lets,especially for leaks with a slower drip frequency.It can accurately count the number of drip leaks with an accuracy rate above 98%,and the relative error of estimating leakage volume flow rate is within 20%.This method can effectively monitor pipeline leaks and provide the reference for making pipeline maintenance decisions.

ship pipelinedrip monitoringmachine visionmean backgroundvirtual coilsleakage flow assessment

姜兴家、刘云志、代英伟、杜太利、李顺琦、邹永久、张跃文、孙培廷

展开 >

大连海事大学轮机工程学院,辽宁大连 116026

徐工集团工程机械股份有限公司高技术装备分公司,江苏徐州 221000

船舶管路 滴漏监测 机器视觉 均值背景 虚拟线圈 泄漏流量估计

国家自然科学基金资助项目

52101400

2024

大连海事大学学报
大连海事大学

大连海事大学学报

CSTPCD北大核心
影响因子:0.469
ISSN:1006-7736
年,卷(期):2024.50(1)
  • 16