摘要
文章设计了工业设备故障音频智能识别系统,并将该技术集成到巡检机器人中,机器人在遂行巡检任务时,可以实时采集巡检区域设备发出的音频数据并传输到机器人内部的边缘计算模块,自动进行模型训练、声学特征提取与对比,出现异常特征及时报警,可实现设备内部故障的识别,解决了通过视觉检测的局限性问题,从而使得检测更加全面准确。
In this paper,it is introduced the audio intelligent recognition system for industrial equipment fault is designed and integrated into the patrol robot.The robot could collect real-time audio data of equipment in inspection area and transmit them to the edge computing module inside itself to automatically carry out model training,extract and compare acoustic feature as well as give an alarm in time when there are abnormal characteristics when it is doing inspection tasks so that the recognition of internal fault of equipment could be realized.As a result,the detection is more comprehensive and accurate through solving the problem of limitation of visual detection.
作者
芦建文
赵楠
Lu Jian-wen;Zhao Nan(Inner Mongolia Xinlian Information Industry Co.,Lud.,Baotou 014010,Inner Mongolia Auonomous Region,China)
出处
《包钢科技》
2022年第3期86-89,共4页
Science & Technology of Baotou Steel
基金
内蒙古自治区应用技术研究与开发资金计划资助项目(2021GG0356)。
关键词
音频识别
巡检机器人
特征提取
智能诊断
audio recognition
patrol robot
feature extraction
intelligent diagnosis
