摘要
无人值守机器人在智能配电房进行巡检、监测采样过程中,需对配电房内的指针式仪表进行自动化识别与判读。对于方型指针式仪表,提出一种基于轮廓特征筛选与直线检测的仪表智能识别方法。在该算法的前处理过程中,针对巡检视角窗口的移动性与多个仪表的分割提出了基于分数体系的算法;在表盘指针识别的图像处理过程中,提出了基于轮廓特征检测与Hough线段识别相结合的方法,并进行聚类估值处理,有效地改进了传统的直线检测方案。利用OpenCV视觉库和C++在计算机中进行算法实现和验证,结果表明,该算法满足精度要求,具有较高的应用价值。
During the inspection and monitoring sampling process, the unattended robots need to automatically recognize and interpret pointer meters in the power distribution room. To recognize the square pointer meters, an intelligent identification method based on contour features selection and line detection is proposed. In the pre-processing of the algorithm, an algorithm based on scores system is proposed for the moving window and the segmentation of multiple meters. Then, in the process of recognition of the pointer on the meter, a method based on contour features detection and Hough Line Transform is presented, which includes clustering method and effectively improves the traditional detecting method. After the algorithm implementation and experimental verification with OpenCV and C++, the result shows that the algorithm meets the accuracy requirements and has high application value.
作者
杨应彬
尹春丽
刘波
王丹
张鼎瑞
梁庆华
YANG Yingbin;YIN Chunli;LIU Bo;WANG Dan;ZHANG Dingrui;LIANG Qinghua(School of Mechanical Engineering,Shanghai Jiao Tong University,Shanghai 200240,China;State Grid Jiangsu Electric Power Co.,Xuzhou Branch,Xuzhou Jiangsu 221000,China)
出处
《机械设计与研究》
CSCD
北大核心
2019年第3期7-11,共5页
Machine Design And Research
基金
国网总部资助科技项目“基于人工智能的多任务协同配电站值守作业机器人研发”资助
