基于卷积记忆神经网络的数字表盘读数识别
|
摘要
针对巡检机器人在执行例检任务时,对数字表盘读数识别精确度低的问题,提出一种基于卷积记忆神经网络的数字表盘抄表算法。对高清摄像机获取的目标图像信息,经过图像轮廓提取算法定位到表盘字符区域,结合卷积神经网络(CNN)和长短期记忆网络(LSTM)模型的特点,提出了卷积记忆神经网络模型(CLSTM),与传统字符识别算法CNN和LSTM相比,此模型既不需要做字符分割,又能够优化特征提取。实验以电表进行测试,结果表明,相比于CNN和LSTM,此模型准确率更高。
To overcome the problem of low accuracy for routine inspection of the patrol robot, a digital dial meter reading algorithm based on convolutional memory neural network is proposed. Firstly, the target image information is obtained by a high-definition camera, and the target image information is located through contour extraction algorithm, Secondly,Combined with the features of the convolution neural network(CNN) and the long and short term memory network(LSTM) model, put forward the convolution memory neural network model(CLSTM), compared with the traditional character recognition algorithm of CNN and LSTM, This model requires neither character segmentation nor feature extraction optimization. The experiment was conducted with electricity meters, and the results showed that this model was more accurate than CNN and LSTM.
To overcome the problem of low accuracy for routine inspection of the patrol robot, a digital dial meter reading algorithm based on convolutional memory neural network is proposed. Firstly, the target image information is obtained by a high-definition camera, and the target image information is located through contour extraction algorithm, Secondly,Combined with the features of the convolution neural network(CNN) and the long and short term memory network(LSTM) model, put forward the convolution memory neural network model(CLSTM), compared with the traditional character recognition algorithm of CNN and LSTM, This model requires neither character segmentation nor feature extraction optimization. The experiment was conducted with electricity meters, and the results showed that this model was more accurate than CNN and LSTM.
引文
[1] 张显明.基于机器视觉的药品标签识别及分拣系统研究[D].长春:吉林大学,2017.
[2] Wang Kai,Boris Babenko,Serge Belongie.End-to-end scene text recognition,Computer Vision (ICCV)[C].2011 IEEE International Conference on IEEE,2011.
[3] Bai Xiang,Yao Cong,Liu Wenyu.Strokelets:A learned multi-scale representation for scene text recognition[J].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2014,25(6):4042-4049.
[4]郑啸,王义真,袁志祥,等.基于卷积记忆神经网络的微博短文本情感分析[J].电子测量与仪器学报,2018,32(3):195-200.
[5] 张国云,向灿群,吴健辉,等.基于改进BP网络的车牌字符识别方法研究[J].计算机应用与软件,2017,34(4):243-248.
[6] 蔡梦倩,张蕾,王炎,等.基于全卷积网络的数字仪表字符识别方法[J].现代计算机(专业版),2018(1):38-43.
[7] Jowati Juhary.Making Sense of e-Learning and Simulations:The Misunderstood Perceptions[J].Procedia-Social and Behavioral Sciences,2012,67(10):229-237.
[8] YiFei Li,Han Cao.Prediction for Tourism Flow based on LSTM Neural Network[J].Procedia Computer Science,2018,129:277-283.
[9] Ha Young Kim,Chang Hyun Won.Forecasting the volatility of stock price index:A hybrid model integrating LSTM with multiple GARCH-type models[J].Expert Systems With Applications,2018,103(8):25-37.
[10] 贾磊.LSTM建模和CTC训练在语音建模技术中的应用[A].中国中文信息学会语音信息专业委员会,第十三届全国人机语音通讯学术会议(NCMMSC2015)论文集[C].中国中文信息学会语音信息专业委员会,2015.
[11] 赵继印,李建坡,郑蕊蕊.自适应学习速率法在变压器故障诊断中的应用[J].吉林大学学报(信息科学版),2008,26(4):415-420.
[12] 杨甲沛,李锵,刘郑,等.基于自适应学习速率的改进型BP算法研究[J].计算机工程与应,2009,45(11):56-58,66.
[2] Wang Kai,Boris Babenko,Serge Belongie.End-to-end scene text recognition,Computer Vision (ICCV)[C].2011 IEEE International Conference on IEEE,2011.
[3] Bai Xiang,Yao Cong,Liu Wenyu.Strokelets:A learned multi-scale representation for scene text recognition[J].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2014,25(6):4042-4049.
[4]郑啸,王义真,袁志祥,等.基于卷积记忆神经网络的微博短文本情感分析[J].电子测量与仪器学报,2018,32(3):195-200.
[5] 张国云,向灿群,吴健辉,等.基于改进BP网络的车牌字符识别方法研究[J].计算机应用与软件,2017,34(4):243-248.
[6] 蔡梦倩,张蕾,王炎,等.基于全卷积网络的数字仪表字符识别方法[J].现代计算机(专业版),2018(1):38-43.
[7] Jowati Juhary.Making Sense of e-Learning and Simulations:The Misunderstood Perceptions[J].Procedia-Social and Behavioral Sciences,2012,67(10):229-237.
[8] YiFei Li,Han Cao.Prediction for Tourism Flow based on LSTM Neural Network[J].Procedia Computer Science,2018,129:277-283.
[9] Ha Young Kim,Chang Hyun Won.Forecasting the volatility of stock price index:A hybrid model integrating LSTM with multiple GARCH-type models[J].Expert Systems With Applications,2018,103(8):25-37.
[10] 贾磊.LSTM建模和CTC训练在语音建模技术中的应用[A].中国中文信息学会语音信息专业委员会,第十三届全国人机语音通讯学术会议(NCMMSC2015)论文集[C].中国中文信息学会语音信息专业委员会,2015.
[11] 赵继印,李建坡,郑蕊蕊.自适应学习速率法在变压器故障诊断中的应用[J].吉林大学学报(信息科学版),2008,26(4):415-420.
[12] 杨甲沛,李锵,刘郑,等.基于自适应学习速率的改进型BP算法研究[J].计算机工程与应,2009,45(11):56-58,66.