摘要
轮胎标识信息是利用不同高度形成的凸凹变化来描述,压印字符与其背景区域对比度较低造成字符识别困难.为了提高检测效率、降低人工成本,研究轮胎上的压印字符识别技术在工业上的应用.通过预处理尽量减少背景带来的噪声干扰,采用改进的随机Hough变换来提取轮胎的同心圆作为感兴趣区域,将基于深度学习的Faster R-CNN方法引入到轮胎字符检测中,显著提高了压印字符识别的准确率和效率.
Tire identification information is described by concave and convex changes formed by different heights;the low contrast between imprinted characters and their background areas makes character recognition difficult.This paper aims to study the application of technology of character recognition imprinted on tires in industry so as to improve the detecting efficiency and cut labor cost.The noise interference caused by background is first minimized by preprocessing,and then the concentric circle of tire is extracted as the region of interest by improved random Hough transform.Finally,the Faster R-CNN method based on deep learning is introduced into tire character detection,which significantly improves the recognition accuracy and efficiency of imprinted characters.
作者
张晨梦
周健
陈颖
ZHANG Chen-meng;ZHOU Jian;CHEN Ying(College of Telecommunications & Information Engineering, Nanjing University of Posts andTelecommunications, Nanjing 210003, China)
出处
《南京工程学院学报（自然科学版）》
2020年第2期50-54,共5页
Journal of Nanjing Institute of Technology(Natural Science Edition)
