为了提高卡尺显示的数字识别效率及识别精度,提出了一种基于改进的穿线法的数显卡尺图像识别算法。传统的穿线法在数字图像识别时对图中数字垂直方向有倾斜的数字图像的识别会频繁出现错误,实用性差,有缺陷,无法在智能制造现场大范围推广。针对这个问题,先采用Radon变换对数字图像的垂直方向倾斜角进行估计,再利用该估计角度的倾斜直线代替传统穿线法竖直直线,进而通过数字图像分割和处理,实现高效率识别。实验结果表明,该算法识别率约为98%,很好地满足了工业测量中大范围高密集度的数字图像识别的使用要求。
Abstract
In order to improve the digital recognition efficiency and precision of caliper display, a digital caliper image recognition algorithm based on improved threading method is proposed. The traditional threading method in the digital image recognition of the digital image with vertical tilt has poor practicality and defects frequently, it cannot be widely promoted in the intelligent manufacturing site. To solve this problem, Radon transform is first used to estimate the vertical inclination a angle of digital image, and then the slant line of the estimated angle is used to replace the traditional vertical line of the threading method, so as to realize efficient recognition through digital image segmentation and processing. Experimental results show that the recognition rate of this algorithm is 98%, which can meet the requirements of high density digital image recognition in industrial measurement.
关键词
计量学;图像识别;机器视觉 /
穿线法 /
数显卡尺
Key words
metrology /
image recognition /
machine vision /
threading method /
digital caliper
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
［1］张竟月, 孙斌, 谢玄达, 等. 基于激光和机器视觉的微量液体容量计量方法［J］. 计量学报,2018,39(4):504-509.
Zhang J Y, Sun B, Xie X D, et al. Micro Liquid Volumetric Method based on Laser and Machine Vision［J］. Acta Metrologic Sinica,2018,39(4):504-509.
［2］马丹丹,王鹏程,刘基宏. 基于机器视觉的条码尺精确测量技术在落纱小车中的应用［J］. 计量学报,2019,40(3):409-415.
Ma D D,Wang P C,Liu J H. Application of Accurate Measurement Technology of Barcode Scale in Doffing Vehicle Based on Machine Vision［J］. Acta Metrologica Sinica,2019,40(3):409-415.
［3］Huang Z Q, Wang C F. Reading recognition of digitaldisplay instrument based on BP neutral network［C］//2008 International Conference on Computer Science and Software Engineering. Wuhan, China, 2008.
［4］Chen H, Wang X, Xu B. Study on digital displayinstrument recognition for substation based on pulse coupledneural network［C］//2016 IEEE International Conference on Information and Automation. Ningbo, China, 2016.
［5］唐轶峻, 申小阳, 朱雯兰, 等. 基于BP神经网络的数显仪表数字字符识别系统［J］. 电测与仪表, 2005, 42(9): 42-45.
Tang Y J, Shen X Y, Zhu W L, et al. Recognition System for Character of Numeral Instrument Dynamic Displayed Based on BP Neural Network［J］. Electical Measurement & Instrumentation, 2005, 42(9): 42-45.
［6］卓磊, 周律, 杨丽红. 基于穿线法的计算器数字识别［J］. 软件工程, 2018, 21(12): 1-3.
Zhuo L, Zhou L, Yang L H. Calculator Number Recognition based on Threading Method［J］. Software Engineering, 2018, 21(12): 1-3.
［7］魏畅然. 基于穿线法的数字识别方法［J］. 科技情报开发与经济, 2010, 20(28): 137-139.
Wei C R. Digital Recognition Method based on Threading Method［J］. Sci-Tech Information Development & Economy, 2010, 20(28): 137-139.
［8］申小阳, 唐轶峻, 姜柏军, 等. 仪表的数字字符识别系统［J］. 仪表技术与传感器, 2005, (6): 55-57.
Shen X Y, Tang Y J, Jiang B J, et al. Numeral Characters Recognition System of Instrument［J］. Instrument Technique and Sensor, 2005, (6): 55-57.
［9］曾科, 高潮, 扶新, 等. 多参数数显仪表的自动识别方法研究［J］. 中国测试, 2018, 44(12): 122-128.
Zeng K, Gao C, Fu X, et al. Automatic recognition method for the digital display instrument with multi-parameters［J］. China Measuremeat & Testing Technology, 2018, 44(12): 122-128.
［10］李倩. 文档图像的二值化算法综述［J］. 中国传媒大学学报(自然科学版), 2008, 15(4): 66-70.
Li Q. Summary of Binarization Algorithms for Document Images［J］. Journal of Communication University of China (Science and Technology), 2008, 15(4): 66-70.
［11］褚慧君, 朱宏辉, 高军. 基于改进的投影检测车牌倾斜校正算法［J］. 物流技术, 2017, 36(1): 89-96.
Chu H J, Zhu H H, Gao J. Improved Tilt Correction Algorithm for Shadowgraph Detection of License Plates［J］. Logistics Technology, 2017, 36(1): 89-96.
［12］王蓉芳, 汪仁煌, 康亚芬. 基于数码管机器视觉的小数点识别［J］. 电子测试, 2008, (11): 17-20.
Wan R F, Wan R F, Kang Y F, et al. Decimal Point Recognition based on Nixie Tube Machine Vision ［J］. Electronic Test, 2008, (11): 17-20.
基金
国家重点研发计划(2018YFF0212702)
PDF(913 KB)
