基于深度学习的X射线焊缝缺陷识别
|
摘要
为了提高缺陷识别效率,提出利用基于深度学习网络进行焊缝缺陷识别。在分析X射线焊缝缺陷图像特征的基础上,构建一种基于模拟视觉感知原理的深度学习网络结构,并对卷积神经网络的卷积模板大小及层数进行了分析,对卷积神经网络隐藏层中2种不同的激活函数进行了实验验证,针对性地提出优化方法。该深度学习神经网络可以避免对焊缝缺陷图像特征的提取,直接判断疑似缺陷图像是否为缺陷。对580张图像进行了实验,结果表明,本文所提方法对SDR图像的识别准确率超过98%,优于传统方法。且所设计系统具有自动学习X射线焊缝缺陷图像中复杂的深度特征的特点,实用性较强。
In order to improve the efficiency of defect recognition,the recognition of X-ray weld defects based on deep learning is proposed.A deep learning network structure based on the principle of simulated visual perception is established based on the analysis of the characteristics of X-ray weld defect images.The convolution template size and layernumber of convolutional neural network are analyzed.Two different activation functions in the hidden layer of convolutional neural network are verified,and an optimization method for them is proposed.To use the deep learning neural network can avoid the step of extracting the features of the weld defect images and directly determine whether the suspected defect image is a defect.Experiments on 580 images show that the recognitionaccuracy of the proposed method to SDR images is over 98%,which shows that the proposed weld defect recognition method based on deep learning network is better than the traditional method.And the designed system has the ability of automatically learning the complex features in X-ray weld defect images,and it has strong practicability.
In order to improve the efficiency of defect recognition,the recognition of X-ray weld defects based on deep learning is proposed.A deep learning network structure based on the principle of simulated visual perception is established based on the analysis of the characteristics of X-ray weld defect images.The convolution template size and layernumber of convolutional neural network are analyzed.Two different activation functions in the hidden layer of convolutional neural network are verified,and an optimization method for them is proposed.To use the deep learning neural network can avoid the step of extracting the features of the weld defect images and directly determine whether the suspected defect image is a defect.Experiments on 580 images show that the recognitionaccuracy of the proposed method to SDR images is over 98%,which shows that the proposed weld defect recognition method based on deep learning network is better than the traditional method.And the designed system has the ability of automatically learning the complex features in X-ray weld defect images,and it has strong practicability.
引文
[1] 张晓光,孙正,胡晓磊,等.射线检测图像中焊缝和缺陷的提取方法[J].焊接学报,2011,32(2):77-80.ZHANG Xiaoguang,SUN Zheng,HU Xiaolei,et al.Extraction of welds and defects in radiographic images[J].Transactions of the China Welding Institution,2011,32(2):77-80.
[2] TANG Yanlong,TONG Ruofeng,TANG Min,et al.Depth incorporating with color improves salient object detection[J].Visual Computer,2016,32(1):111-121.
[3] 李勇,高炜欣,汤楠,等.基于压缩感知的X射线螺旋焊管焊缝缺陷检测[J].焊接技术,2013,42(2):51-55,76.LI Yong,GAO Weixin,TANG Nan,et al.Defect detection of X-ray spiral welded pipe weld based on compressed sensing[J].Welding Technology,2013,42(2):51-55,76.
[4] 高炜欣,胡玉衡,武晓朦.基于压缩传感技术的埋弧焊X射线焊缝图像缺陷检测[J].焊接学报,2015,36(11):85-88,117.GAO Weixin,HU Yuheng,WU Xiaomeng.Image defect detection of submerged arc welding X-ray weld based on compression sensing technology[J].Transactions of the China Welding Institution,2015,36(11):85-88,117.
[5] 王家晨,王新房.基于背景重构X射线钢管焊缝缺陷检测方法[J].计算机系统应用,2018,27(2):245-249.WANG Jiachen,WANG Xinfang.Detection method of weld defect in X-ray steel tube based on backgroundreconstruction[J].Computer Systems & Applications,2018,27(2):245-249.
[6] 戴忠晨,孟宪伟,付宁宁,等.铝合金搅拌摩擦焊焊缝的超声相控阵扫描图像处理算法[J].焊接技术,2018,47(1):55-57.DAI Zhongchen,MENG Xianwei,FU Ningning,et al.Ultrasonic phased array scanning image processing algorithm for friction stir welding of aluminum alloy[J].Welding Technology,2018,47(1):55-57.
[7] 朱红秀,刘欢,李宏远,等.基于优化RBF神经网络的管道缺陷量化分析方法[J].仪表技术与传感器,2016(2):83-86.ZHU Hongxiu,LIU Huan,LI Hongyuan,et al.Quantitative analysis method of pipeline defects based on optimized RBF neural network[J].Instrument Technique and Sensor,2016(2):83-86.
[8] LIM T Y,RATNAM M M,KHALID M A.Automatic classification of weld defects using simulated data and an MLP neural network[J].Insight,2007,49(3):154-159.
[9] 王雒瑶,高炜欣,王欣.一种基于SVM及LE降维的X射线焊缝缺陷分类算法研究[J].西安石油大学学报(自然科学版),2017,32(5):96-101,106.WANG Luoyao,GAO Weixin,WANG Xin.A classification algorithm for X-ray weld defects based on SVM and LE dimensionality reduction[J].Journal of Xi'an Shiyou University(Natural Science Edition),2017,32(5):96-101,106.
[10] 王征,王欣,高炜欣,等.图像降维下的埋弧焊缺陷自动识别算法及框架[J].焊接,2016(9):12-16,72.WANG Zheng,WANG Xin,GAO Weixin,et al.Automatic identification algorithm and framework for submerged arc welding defects under image dimensionality reduction[J].Welding & Joining,2016(9):12-16,72.
[11] 孙林,杨世元,吴德会.X射线底片焊缝缺陷的支持向量机识别方法[J].应用科学学报,2008,26(4):418-424.SUN Lin,YANG Shiyuan,WU Dehui.Support vector machine identification method for X-ray film weld defects[J].Journal of Applied Sciences,2008,26(4):418-424.
[12] 武晓朦,高炜欣,袁磊,等.SVMs在油气管道焊缝缺陷检测中的应用[J].西安石油大学学报(自然科学版),2011,26(6):97-101,112.WU Xiaomeng,GAO Weixin,YUAN Lei,et al.Application of SVMs in the detection of weld defects in oil and gas pipelines[J].Journal of Xi'an Shiyou University(Natural Science Edition),2011,26(6):97-101,112.
[13] VALAVANIS I,KOSMOPOULOS D.Multiclass defect detection and classification in weld radiographic images using geometric and texture features[J].Expert Systems with Applications,2010,37(12):7606-7614.
[14] 张晓光,林家骏.基于模糊神经网络的焊缝缺陷识别方法的研究[J].中国矿业大学学报,2003,32(1):92-95.ZHANG Xiaoguang,LIN Jiajun.Weld defects distinguishing method based on fuzzy neural networks[J].Journal of China University of Mining & Technology,2003,32(1):92-95.
[15] HERNANDEZ S,SAEZ D,MERY D.Neuro-fuzzy method for automated defect detection in aluminium castings[C].1st International Conference on Image Analysis and Recognition,September 29 - October 1,2004.Springer Verlag:826-833.
[16] HINTON G E,SALAKHUTDINOV R R.Reducing the dimensionality of data with neural networks[J].Science,2006,313(5786):504-507.
[2] TANG Yanlong,TONG Ruofeng,TANG Min,et al.Depth incorporating with color improves salient object detection[J].Visual Computer,2016,32(1):111-121.
[3] 李勇,高炜欣,汤楠,等.基于压缩感知的X射线螺旋焊管焊缝缺陷检测[J].焊接技术,2013,42(2):51-55,76.LI Yong,GAO Weixin,TANG Nan,et al.Defect detection of X-ray spiral welded pipe weld based on compressed sensing[J].Welding Technology,2013,42(2):51-55,76.
[4] 高炜欣,胡玉衡,武晓朦.基于压缩传感技术的埋弧焊X射线焊缝图像缺陷检测[J].焊接学报,2015,36(11):85-88,117.GAO Weixin,HU Yuheng,WU Xiaomeng.Image defect detection of submerged arc welding X-ray weld based on compression sensing technology[J].Transactions of the China Welding Institution,2015,36(11):85-88,117.
[5] 王家晨,王新房.基于背景重构X射线钢管焊缝缺陷检测方法[J].计算机系统应用,2018,27(2):245-249.WANG Jiachen,WANG Xinfang.Detection method of weld defect in X-ray steel tube based on backgroundreconstruction[J].Computer Systems & Applications,2018,27(2):245-249.
[6] 戴忠晨,孟宪伟,付宁宁,等.铝合金搅拌摩擦焊焊缝的超声相控阵扫描图像处理算法[J].焊接技术,2018,47(1):55-57.DAI Zhongchen,MENG Xianwei,FU Ningning,et al.Ultrasonic phased array scanning image processing algorithm for friction stir welding of aluminum alloy[J].Welding Technology,2018,47(1):55-57.
[7] 朱红秀,刘欢,李宏远,等.基于优化RBF神经网络的管道缺陷量化分析方法[J].仪表技术与传感器,2016(2):83-86.ZHU Hongxiu,LIU Huan,LI Hongyuan,et al.Quantitative analysis method of pipeline defects based on optimized RBF neural network[J].Instrument Technique and Sensor,2016(2):83-86.
[8] LIM T Y,RATNAM M M,KHALID M A.Automatic classification of weld defects using simulated data and an MLP neural network[J].Insight,2007,49(3):154-159.
[9] 王雒瑶,高炜欣,王欣.一种基于SVM及LE降维的X射线焊缝缺陷分类算法研究[J].西安石油大学学报(自然科学版),2017,32(5):96-101,106.WANG Luoyao,GAO Weixin,WANG Xin.A classification algorithm for X-ray weld defects based on SVM and LE dimensionality reduction[J].Journal of Xi'an Shiyou University(Natural Science Edition),2017,32(5):96-101,106.
[10] 王征,王欣,高炜欣,等.图像降维下的埋弧焊缺陷自动识别算法及框架[J].焊接,2016(9):12-16,72.WANG Zheng,WANG Xin,GAO Weixin,et al.Automatic identification algorithm and framework for submerged arc welding defects under image dimensionality reduction[J].Welding & Joining,2016(9):12-16,72.
[11] 孙林,杨世元,吴德会.X射线底片焊缝缺陷的支持向量机识别方法[J].应用科学学报,2008,26(4):418-424.SUN Lin,YANG Shiyuan,WU Dehui.Support vector machine identification method for X-ray film weld defects[J].Journal of Applied Sciences,2008,26(4):418-424.
[12] 武晓朦,高炜欣,袁磊,等.SVMs在油气管道焊缝缺陷检测中的应用[J].西安石油大学学报(自然科学版),2011,26(6):97-101,112.WU Xiaomeng,GAO Weixin,YUAN Lei,et al.Application of SVMs in the detection of weld defects in oil and gas pipelines[J].Journal of Xi'an Shiyou University(Natural Science Edition),2011,26(6):97-101,112.
[13] VALAVANIS I,KOSMOPOULOS D.Multiclass defect detection and classification in weld radiographic images using geometric and texture features[J].Expert Systems with Applications,2010,37(12):7606-7614.
[14] 张晓光,林家骏.基于模糊神经网络的焊缝缺陷识别方法的研究[J].中国矿业大学学报,2003,32(1):92-95.ZHANG Xiaoguang,LIN Jiajun.Weld defects distinguishing method based on fuzzy neural networks[J].Journal of China University of Mining & Technology,2003,32(1):92-95.
[15] HERNANDEZ S,SAEZ D,MERY D.Neuro-fuzzy method for automated defect detection in aluminium castings[C].1st International Conference on Image Analysis and Recognition,September 29 - October 1,2004.Springer Verlag:826-833.
[16] HINTON G E,SALAKHUTDINOV R R.Reducing the dimensionality of data with neural networks[J].Science,2006,313(5786):504-507.