基于改进水平集算法的锅炉承压视觉预警技术  

Research on boiler pressure warning technology based on intelligent vision

作　　者：张苏颖 ZHANG Suying(Jiangsu Naning Higher Vocational School of Engineering,Nanjing 211135)

机构地区：[1]江苏省南京工程高等职业学校,南京211135

出　　处：《工业安全与环保》2022年第2期32-36,共5页Industrial Safety and Environmental Protection

摘　　要：传统锅炉承压预警技术主要依靠人工巡查发现问题,在检测过程中受工作人员自身操作不稳定因素影响,存在故障预警实时性差、盲点多等问题,达不到相应的预警要求。针对此问题,提出基于改进采用水平集算法的锅炉承压预警技术。利用图像处理技术从图像中分割出锅炉承压关键部件,统计关键部件灰度值,并利用灰度值设定特征图像及初始轮廓,设计采用水平集算法实现锅炉承压关键部件的自动分割。提取分割获得的关键部件目标特征向量,设计支持向量机分类器完成对算法的改进,根据识别结果及温度信息,对锅炉承压特种设备的故障等级做出判断,进行故障预警。实验结果表明,该技术可实现锅炉承压设备自动故障预警,具有实时性。In view of the traditional boiler pressure early warning technology,it mainly relies on manual inspection to find problems,which is affected by the unstable factors of staff operation in the detection process,so there are some problems such as poor real-time fault early warning,many blind spots and so on,which can not meet the corresponding early warning requirements.To solve this problem,a pressure early warning technology for boiler based on improved level set algorithm is proposed.The image processing technology is used to segment the key components of the boiler pressure and count the gray value of the key components.The gray value is used to set the feature image and the initial contour and the level set algorithm is designed to realize the automatic segmentation of the key components of the boiler pressure.Target feature vectors of key components obtained by segmentation are extracted,and support vector machine classifier is designed to improve the algorithm.According to the recognition results and temperature information,the fault grade of boiler pressure special equipment is judged and the fault early warning is carried out.The experimental results show that the technology can realize automatic fault warning of boiler pressure equipment and has real-time performance.

关 键 词：智能视觉 锅炉承压 预警 支持向量机分类器 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]