期刊文献+

基于改进U-Net网络的激光条纹提取方法

Laser Stripe Extraction Method Based on Improved U-Net Network
原文传递
导出
摘要 为解决在电弧光、烟雾和大量火花飞溅等强噪声的背景下,使用传统的基于图像处理的激光条纹提取方法灵活性和鲁棒性较差的问题,提出了一种基于全卷积神经网络(DB-U-Net)的激光条纹提取方法。实验表明,通过在模型主干网络引入密集残差块(DB)和注意力机制,提高了模型对焊缝图像的全局信息提取能力,模型综合性能AUC_(PR)从0.891提升到0.924;并在此基础上采用层级深度监督训练模式结合多层特征级联输出,融合低层与高层特征信息,减少了因多次上、下采样及深层网络卷积操作造成的信息丢失,模型综合性能AUC_(PR)从0.924提升到0.932。利用所提出的模型网络对强噪声图像去噪处理后进行激光条纹中心线提取的位置误差达到0.50 pixel,证明该方法对焊接过程中强烈的图像噪声干扰具有较高的检测精度和鲁棒性。 Under the background of strong noise such as arc light,smoke and a lot of spark splash,conventional laser stripe extraction method based on image processing have the shortcomings of poor flexibility and robustness.In this paper,a method of laser stripe extraction based on full convolution neural network(DB-U-Net)is proposed.The experiments show that by introducing dense residual block(DB)and attention mechanism into the backbone network of the model,the global information extraction ability of the model is improved,and the comprehensive performance AUC_(PR)of the model is increased from 0.891 to 0.924.On this basis,the multi-level deep supervision training mode combined with multi-layer feature cascading output is used to integrate the low-level and high-level feature information,which reduces the information loss caused by multiple up and down-sampling and deep network convolution operation.The comprehensive performance AUC_(PR)of the model is increased from 0.924 to 0.932.By using the proposed model network,the position error of laser stripe centerline extraction after de-noising the strong noise image is up to 0.50 pixel,which proves that the method has high detection accuracy and robustness against the strong noise interference in the welding process.
作者 李雪梅 郭义华 张鑫 何志威 常浩 Li Xuemei;Guo Yihua;Zhang Xin;He Zhiwei;Chang Hao(School of Mechanical and Electrical Engineering,Guilin University of Electronic Technology,Guilin 541004,Guangxi,China)
出处 《应用激光》 CSCD 北大核心 2024年第9期133-146,共14页 Applied Laser
基金 广西研究生教育创新计划资助项目(YCSW2022276)。
关键词 图像去噪 激光条纹提取 全卷积神经网络 特征级联 注意力机制 image denoising laser stripe extraction full convolutional neural network feature cascade attention mechanism
  • 引文网络
  • 相关文献

参考文献4

二级参考文献27

共引文献17

;
使用帮助 返回顶部