摘要
针对精密指针式仪表读数算法对自动读数的自适应性差问题,提出一种基于SVM的精密指针式仪表自动读数方法。该方法首先对二值化处理后的仪表图像通过面积形态学提取刻度并做极坐标变化,再利用面积形态学特征提取刻度与指针位置,最后利用SVM识别刻度对应的数值进行读数判读。实验表明,自动读数结果与实际数值相比误差均小于0.1%,算法稳定可靠。
Aiming at the problem of adaptivity in autoreading method for precision pointer meter, an auto-reading method for precision pointer meter based on support vector machine(SVM) is proposed. Firstly, the proposed approach extract the scale by area of morphological features of binary meter image, and polar transformation image is derived by information of arc of the scale. Secondly, the scale and index are extracted in polar transformation image using max-area information of binary polar transformation meter image. Finally, the digits of the scale values are recognized by SVM, and the meter reading is calculated by the relative position between scale and index. Experimental results show that the relative error between meter auto-readings and actual readings is less than 0.1% and the method is reliable.
引文
[1]MOGHADDAM R F,CHERIET M.A Multi-scale Framework for Adaptive Binarization of Degraded Document Images[J].Pattern Recognition,2010,43(6):2186-2198.
[2]李治玮,郭戈.一种新型指针仪表识别方法研究[J].微计算机信息,2007,(31):113-114,126.
[3]张冀,王俊宏,尉迟明,等.基于计算机视觉的汽车仪表指针检测方法[J].计算机工程与科学,2013,35(3):134-139.
[4]施健,张冬,何建国,等.一种指针式化工仪表的远程抄表设计方法[J].自动化仪表,2014,35(5):77-79.
[5]房桦,明志强,周云峰,等.一种适用于变电站巡检机器人的仪表识别算法[J].自动化与仪表,2013,28(5):10-14.
[6]温和,滕召胜,杨圣洁,等.基于计算机视觉的指针式仪表智能判读方法[J].仪器仪表学报,2007,28(7):1234-1239.
[7]王三武,戴亚文.多指针水表自动识别系统[J].仪器仪表学报,2005,26(11):1178-1181.
[8]Bovik A L.Handbook of Image and Video Processing[M].New York:Academic Press,2000.
[9]Astorino A,Gorgone E,Gaudioso M,et al.Data Preprocessing in Semi-supervised SVM Classification[J].Optimization,2011,60(1-2):143-151.
[10]蒋华,戚玉顺.基于球结构SVM的多标签分类[J].计算机工程,2013,39(1):294-297.
[11]奉国和.SVM分类核函数及参数选择比较[J].计算机工程与应用,2011,47(3):123-124.