基于ELSD的指针式仪表自动识别算法研究
|
摘要
为实现指针式仪表自动识别,本文提出一种基于直线椭圆检测器(ELSD)的指针仪表自动识别策略。整个策略包括刻度线粗识别、仪表圆弧边界拟合、刻度线重识别和排序、仪表指针识别以及仪表示数计算等5个模块。首先,使用本文设计的自动迭代算法将检测到的线段匹配成对,并结合K-means聚类算法得到粗识别的刻度线;接着,根据粗识别出的刻度线与检测出的仪表圆弧位置关系确定仪表圆弧边界;然后,根据仪表刻度线与圆弧边界相交重新确定刻度线并进行插值补全、排序,构建出整个仪表刻度线坐标系;最后,根据仪表指针自身结构特征得到指针在坐标系中的位置并计算出仪表示数。通过实验结果可知,本文提出的算法可减少对模板图像和人工标定的依赖,在保证准确率的情况下能简化并加快整个仪表自动识别过程。
In order to realize the automatic identification of pointer meters, an automatic recognition strategy for pointer meters based on ellipse and line segment detector(ELSD) is proposed. The whole strategy includes five modules: tick mark rough recognition, instrument arc boundary fitting, tick mark re-identification and sorting, instrument pointer recognition and instrument representation number calculation. Firstly, the automatic iterative algorithm designed in this paper is used to match the detected line segments into pairs, and the K-means clustering algorithm is used to obtain the rough recognized tick marks. Then, according to the rough recognized tick marks and the detected instrument arc positions, the relationship determines the arc boundary of the instrument. Thirdly, according to the intersection of the meter tick and the arc boundary, the tick mark is re-determined and the interpolation is completed and sorted to construct the entire instrument tick coordinate system. Finally, the pointer is obtained according to the structural characteristics of the meter pointer itself. The position in the coordinate system and the number of representations is calculated. The experiment results show that the proposed algorithm can reduce the dependence on template image and manual calibration, and simplify and speed up the automatic identification process of the whole instrument while ensuring its accuracy.
In order to realize the automatic identification of pointer meters, an automatic recognition strategy for pointer meters based on ellipse and line segment detector(ELSD) is proposed. The whole strategy includes five modules: tick mark rough recognition, instrument arc boundary fitting, tick mark re-identification and sorting, instrument pointer recognition and instrument representation number calculation. Firstly, the automatic iterative algorithm designed in this paper is used to match the detected line segments into pairs, and the K-means clustering algorithm is used to obtain the rough recognized tick marks. Then, according to the rough recognized tick marks and the detected instrument arc positions, the relationship determines the arc boundary of the instrument. Thirdly, according to the intersection of the meter tick and the arc boundary, the tick mark is re-determined and the interpolation is completed and sorted to construct the entire instrument tick coordinate system. Finally, the pointer is obtained according to the structural characteristics of the meter pointer itself. The position in the coordinate system and the number of representations is calculated. The experiment results show that the proposed algorithm can reduce the dependence on template image and manual calibration, and simplify and speed up the automatic identification process of the whole instrument while ensuring its accuracy.
引文
[ 1] 李伟,任青青,胡艳侠,等.一种复杂指针式仪表的读数自动识别算法[J].计算机技术与发展,2017,27(3):201-204
[ 2] 许丽,石伟,方甜,等.巡检机器人中的指针式仪表读数识别系统[J].仪器仪表学报,2017,38(7):1782-1790
[ 3] 新勤,周睿.船用指针式仪表自动识别方法研究[J].计测技术,2017,37(4):49-52
[ 4] 何智杰,张彬,金连文.高精度指针仪表自动读数识别方法[J].计算机辅助工程,2006,15(3):9-12
[ 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] Matas J,Kittler J,Galambos C.Robust detection of lines using the progressive probabilistic Hough transform[J].Computer Vision & Image Understanding,2000,78(1):119-137
[ 7] Ye W,Hu J D,Zou W B.An auto-recognition system for pointer-meter based on TLS analytic resolution[C].In:Proceedings of the International Conference on Networking and Digital Society,Guiyang,China,2009.83-86
[ 8] 陈明晶,方源敏,陈杰.最小二乘法和迭代法圆曲线拟合[J].测绘科学,2016,41(1):194-197
[ 9] Grompone v G R,Jakubowicz J,Morel J M,et al.LSD:a fast line segment detector with a false detection control[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2010,32(4):722-732
[10] Pǎtrǎucean V,Gurdjos P,Gioi R G V.A parameterless line segment and elliptical arc detector with enhanced ellipse fitting[C].In:Proceedings of the European Conference on Computer Vision,Florence,Italy,2012.97-113
[ 2] 许丽,石伟,方甜,等.巡检机器人中的指针式仪表读数识别系统[J].仪器仪表学报,2017,38(7):1782-1790
[ 3] 新勤,周睿.船用指针式仪表自动识别方法研究[J].计测技术,2017,37(4):49-52
[ 4] 何智杰,张彬,金连文.高精度指针仪表自动读数识别方法[J].计算机辅助工程,2006,15(3):9-12
[ 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] Matas J,Kittler J,Galambos C.Robust detection of lines using the progressive probabilistic Hough transform[J].Computer Vision & Image Understanding,2000,78(1):119-137
[ 7] Ye W,Hu J D,Zou W B.An auto-recognition system for pointer-meter based on TLS analytic resolution[C].In:Proceedings of the International Conference on Networking and Digital Society,Guiyang,China,2009.83-86
[ 8] 陈明晶,方源敏,陈杰.最小二乘法和迭代法圆曲线拟合[J].测绘科学,2016,41(1):194-197
[ 9] Grompone v G R,Jakubowicz J,Morel J M,et al.LSD:a fast line segment detector with a false detection control[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2010,32(4):722-732
[10] Pǎtrǎucean V,Gurdjos P,Gioi R G V.A parameterless line segment and elliptical arc detector with enhanced ellipse fitting[C].In:Proceedings of the European Conference on Computer Vision,Florence,Italy,2012.97-113