摘要
论文在对已有的各种使用图像处理手段进行仪表指针识别方案进行研究的基础上,根据工控环境中指针式仪表所具有特点,设计了基于多分辨率处理、霍夫变换的快速、准确的指针识别算法,该算法是已有识别方案的改进与融合。为了对算法性能进行测试,论文在所提算法的基础上,设计了简单并且容易部署的指针识别系统。识别系统的运行数据证明,论文所提算法可在资源有限的嵌入式系统中对工控环境中的指针式仪表读数进行快速而准确的识别。
This paper firstly studied some old digital image processing style pointer meter analysis algorithms and industrial environment pointer instrument＇s features. And then proposed a new pointer meter analysis algorithm. Proposed algorithm is a merger and an improvement of these old algorithms, the core idea is multi-resolution processing and Hough transform. In order to test the proposed algorithm, this paper designed a pointer meter analyzer system which could be easily deployed in industrial environment. System run data show that proposed algorithm could be applied in embedded system to do pointer instrument analysis in a short time.
出处
《软件》
2012年第8期138-143,共6页
Software
关键词
模式识别
指针识别
嵌入式系统
霍夫变换
多分辨率处理
pattern recognition
pointer meter analyzer
embedded system
hough transform
multi-resolution processing
