摘要
为了快速准确获得工业锅炉汽包水位实时监测值,采用了基于FPGA的图像预处理和数字识别算法,包括双线性插值、最小的误分类误差阈值分割法、Do G滤波器(高斯差分滤波器)及缝隙码,并提出一种新算法—横竖端点寻线法(HVPL),对采集图像进行畸变校正、分割、提取及数字识别,从而实现了水位计图像数字化。实验结果表明,采用基于逻辑电路的快速算法实现对水位计图像的机器识别具有较好实时性、鲁棒性以及较高识别率。
To fast and accurately obtain the real-time monitoring value of steam drum water level for industrial boiler, image preprocessing and digital recognition algorithms based on FPGA are adopted, including bilinear interpolation, DoG （difference of Gaussian） filter, smallest misclassification error threshold method, crack chain codes and HVPL （ searching endpoints and intersection line segments by horizontal and vertical scanning） method for image correction, segmentation, extraction and digital identification, thus achieves the gauge image digitization Experiment results show that the fast algorithms realizing intelligent reading of image based on logic circuits have good real-time performance, robustness and higher recognition rate.
出处
《控制工程》
CSCD
北大核心
2015年第3期413-417,共5页
Control Engineering of China
基金
云南省中青年学术和技术带头人后备人才项目(2012HB011)
昆明理工大学学科方向建设项目(14078212)
作者简介
孙斌（1990-），男，湖南常德人，研究生，主要研究方向为模式识别与嵌入式系统等。
