基于图像处理的加工工艺信号数据自动提取方法
|
摘要
在工业生产过程中,大多对多个加工工艺信号数据进行连续采集,而后续信号分析和处理时却往往只需要某一段加工工艺的信号数据,因此需要从连续多个加工工艺信号数据中提取单个加工工艺信号数据。基于图像处理,采用二维卷积计算方法,提出一种加工工艺信号数据提取方法。通过建立加工工艺信号数据模板图像库,应用图像匹配方法将模板图像库与实际加工工艺信号数据图像进行匹配,进而选取最佳模板图像。采用二维卷积计算方法对最佳模板图像与实际加工工艺信号数据图像进行定位,以实现多个不同加工工艺信号数据的同时提取,并按时间顺序自动进行分类存储。通过实例验证,这一提取方法的精度满足后续分析及处理的要求。
In the industrial process, multiple processing signal data are continuously collected in most cases, but the subsequent signal analysis and processing sometimes only need the signal data of a certain process, so it is necessary to extract a single processing signal data from a series of multiple processing signal data. Based on image processing, a 2 D convolution calculation method was adopted to put forward a method for extraction of processing signal data. By establishing a template image library of processing signal data, the image matching method was used to match the template image library with the actual processing signal data image, and then select the best template image. The 2 D convolution calculation method was used to locate the optimal template image and the actual processing signal data image to realize simultaneous extraction of multiple different processing signal data while performing classification and storage in chronological order automatically. By way of example verification, the accuracy of this extraction method can meet the requirements of subsequent analysis and processing.
In the industrial process, multiple processing signal data are continuously collected in most cases, but the subsequent signal analysis and processing sometimes only need the signal data of a certain process, so it is necessary to extract a single processing signal data from a series of multiple processing signal data. Based on image processing, a 2 D convolution calculation method was adopted to put forward a method for extraction of processing signal data. By establishing a template image library of processing signal data, the image matching method was used to match the template image library with the actual processing signal data image, and then select the best template image. The 2 D convolution calculation method was used to locate the optimal template image and the actual processing signal data image to realize simultaneous extraction of multiple different processing signal data while performing classification and storage in chronological order automatically. By way of example verification, the accuracy of this extraction method can meet the requirements of subsequent analysis and processing.
引文
[1]ROZEWSKI C M.A Method for Measuring and Reporting Manual Data Extraction Reliability[J].Computer Methods and Programs in Biomedicine,1993,41(1):17-31.
[2]陈龙聪,刘改琴,邓世雄,等.从波形图中快速提取信号数据序列方法的研究[J].激光杂志,2012,33(5):46-48.
[3]赵丽娜,李五一,段垚奇,等.飞行实测信号中特征数据的模糊提取[J].测试技术学报,2015,29(5):379-384.
[4]MITRA S,MITRA M.An Automated Data Extraction System from 12 Lead ECG Images[J].Computer Methods and Programs in Biomedicine,2003,71(1):33-38.
[5]杨喜东.基于频域分析的图像匹配定位算法的研究[J].通信电源技术,2012,29(4):20-22,43.
[6]戴涛,朱长仁,胡树平.图像匹配技术综述[J].数字技术与应用,2012(3):174-175,177.
[7]程欣.基于深度学习的图像目标定位识别研究[D].成都:电子科技大学,2016.
[8]梅鸿翔.一种基于图像坐标检测的目标定位方法[J].计算机与数字工程,2016,44(3):438-444.
[9]张辉,胡广书.基于二维卷积的图像插值实时硬件实现[J]清华大学学报(自然科学版),2007,47(6):885-888.
[10]余小游,孙广富,卢焕章.数字图像预处理技术及其硬件实现[J].电子技术应用,1997(11):35-36,38.
[2]陈龙聪,刘改琴,邓世雄,等.从波形图中快速提取信号数据序列方法的研究[J].激光杂志,2012,33(5):46-48.
[3]赵丽娜,李五一,段垚奇,等.飞行实测信号中特征数据的模糊提取[J].测试技术学报,2015,29(5):379-384.
[4]MITRA S,MITRA M.An Automated Data Extraction System from 12 Lead ECG Images[J].Computer Methods and Programs in Biomedicine,2003,71(1):33-38.
[5]杨喜东.基于频域分析的图像匹配定位算法的研究[J].通信电源技术,2012,29(4):20-22,43.
[6]戴涛,朱长仁,胡树平.图像匹配技术综述[J].数字技术与应用,2012(3):174-175,177.
[7]程欣.基于深度学习的图像目标定位识别研究[D].成都:电子科技大学,2016.
[8]梅鸿翔.一种基于图像坐标检测的目标定位方法[J].计算机与数字工程,2016,44(3):438-444.
[9]张辉,胡广书.基于二维卷积的图像插值实时硬件实现[J]清华大学学报(自然科学版),2007,47(6):885-888.
[10]余小游,孙广富,卢焕章.数字图像预处理技术及其硬件实现[J].电子技术应用,1997(11):35-36,38.