巡检机器人中的指针式仪表读数识别系统

发布时间：2018-02-01 07:51

本文关键词： 仪表读数识别指针提取指针夹角计算最大类间方差 Hough变换　出处：《仪器仪表学报》2017年07期 　论文类型：期刊论文

【摘要】：为了解决智能巡检机器人仪表读数识别中易受光照变化影响、识别精度不高等问题,结合高压变电场中常见指针区域的图像特点,建立了指针式仪表读数高精度识别系统。鉴于巡检机器人的室外工作环境,提出了迭代最大类间方差算法,实现了多种光照条件下仪表图像的指针区域提取。经过分析指针转动和图像特性,提出了基于Hough变换的指针角度识别,推导出指针角度与仪表读数的函数关系。该算法较传统Hough变换角度提取法,增加了指针中心线通过表计中心等约束条件,提高了指针角度提取的精度,降低了搜索数据量和搜索时间。通过大量实验验证所建立的表计读数识别系统可实现室外各种光照条件下表计读数识别,获得95%以上的正确识别率。多组鲁棒性实验分析表明,该系统对光照条件、指针宽度、表盘干扰、拍摄角度(不产生指针阴影)具有较好的鲁棒性,但由于拍摄角度而产生指针阴影时,会引起较大的指针中心线提取和角度计算偏差,从而降低仪表读数识别精度。
[Abstract]:In order to solve the problems such as easy to be affected by illumination change and low recognition precision in intelligent inspection robot instrument reading recognition, the image characteristics of common pointer area in high-voltage substation field are combined. A high precision recognition system for the reading of an exponential instrument is established. In view of the outdoor working environment of the robot, an iterative maximum inter-class variance algorithm is proposed. After analyzing the rotation of the pointer and the characteristics of the image, the pointer angle recognition based on Hough transform is proposed. The function relation between pointer angle and meter reading is deduced. Compared with the traditional Hough transform angle extraction method, the algorithm increases the constraint condition of pointer centerline through the center of meter, and improves the precision of pointer angle extraction. Through a large number of experiments to verify the establishment of the meter reading recognition system can achieve outdoor various light conditions meter reading recognition. The experimental results show that the system is robust to illumination conditions, pointer width, dial interference and shooting angle (no pointer shadow). However, when the pointer shadow is produced because of the shooting angle, it will lead to a large pointer centerline extraction and angle calculation deviation, thus reducing the precision of meter reading recognition.
【作者单位】：华北水利水电大学信息工程学院;许继电源有限公司;
【基金】：国家自然科学基金(51609086)项目资助
【分类号】：TP242;TP391.41
【正文快照】： 1引言近年来随着国家电网无人值守变电站的大力推广,智能巡检机器人将逐步代替人工完成变电站设备的巡检[1-3]。巡检机器人携带着可见光、红外、声音等各类传感器,采集电力仪表设备的声、像信息,采用图像处理和模式识别技术完成对各类设备状态的自动识别[4-6]。变电站待巡检的

【相似文献】