摘要
为了对数显仪表数字识别系统进行改进和优化设计,提出了一种新的识别和研究方法,并设计了一个七段数码管和液晶屏数显示仪表的实时数字自动识别系统。先对仪表数字图像二值化、去除噪声等预处理算法进行了研究,利用各数字背景四连通区域对象的个数和标识矩阵元素的分布来构造识别特征从而对数字进行分类识别。采用基于模型编程的方法结合系统仿真软件Simulink建立仿真模型。该系统识别率达98.6%以上,实验结果表明了该模型和方法的有效性和实用性。
To improve and optimize the numeral instrument digital recognition system, a new recognition and research method is proposed. Based on which a real-time automatic recognition system for digits on numeral meter displayed by seven-segment LED and LCD is designed. Firstly, some preprocessing algorithms are studied, such as image binarization, noise removing. Then, the recognition features are structured to recognize the digital images, which mainly include the number of labeled four-connected domain objects of the figure background and label matrix element distribution features. Finally, the system simulation model is established and simulated based on model-oriented technique and dynamic simulation software Simulink. Practical experiments show that the recognition rate is higher than 98.6%, which tells that the model and method are effective and practical.
出处
《计算机工程与设计》
CSCD
北大核心
2010年第1期214-217,共4页
Computer Engineering and Design
关键词
数字识别
数显仪表
自适应阈值二值化
去噪算法
拓扑结构
仿真
digitalrecognition
numeralinstrument
adaptive threshold binarization
noise removing algorithm
topologicalstructure
simulation
