Please wait a minute...
IET Cyber-Systems and Robotics
IET Cyber-Systems and Robotics  2020, Vol. 2 Issue (1): 43-52    DOI: 10.1049/iet-csr.2019.0040
    
基于深度神经网络的滤波增强层析PIV重建
Jiaming Liang1, Shengze Cai1, Chao Xu1, Jian Chu1,2
1 Institute of Cyber-Systems and Control, College of Control Science and Engineering, Zhejiang University, Hangzhou, People's Republic of China
2 Ningbo Industrial Internet Institute (NIII), Ningbo, Zhejiang, People's Republic of China
Filtering enhanced tomographic PIV reconstruction based on deep neural networks
Jiaming Liang1, Shengze Cai1, Chao Xu1, Jian Chu1,2
1 Institute of Cyber-Systems and Control, College of Control Science and Engineering, Zhejiang University, Hangzhou, People's Republic of China
2 Ningbo Industrial Internet Institute (NIII), Ningbo, Zhejiang, People's Republic of China
 全文: PDF 
摘要： 近年来，层析粒子图像测速(Tomo-PIV)已成功地应用于三维流场的测量。该技术高度依赖于基于多台摄像机在不同视角的图像重建空间粒子分布的重建技术。作为最受欢迎的重建方法，乘代数重建法（MART）在稀疏粒子分布重建方面具有较高的计算速度和较高的精度。然而，在稠密粒子分布的情况下，其重建精度并不令人满意。针对这一问题，本文提出了一种对称编码-解码全卷积网络来提高MART的重构质量。该神经网络的输入是MART方法重建的粒子场，输出是相同分辨率的重建图像。数值计算结果表明，经过训练的神经网络可以有效地细化模糊或不规则的粒子。大部分的虚粒子也可以用这种过滤方法去除。同时，在不增加计算量的前提下，重建精度可提高10%以上。实验结果表明，训练后的神经网络能提供令人满意的重建效果以及改进的速度场。
Abstract: Tomographic particle image velocimetry (Tomo-PIV) has been successfully applied in measuring three-dimensional (3D) flow field in recent years. Such technology highly relies on the reconstruction technique which provides the spatial particle distribution by using images from multiple cameras at different viewing angles. As the most popular reconstruction method, the multiplicative algebraic reconstruction technique (MART) has advantages in high computational speed and high accuracy for low particle seeding reconstruction. However, the accuracy is not satisfactory in the case of dense particle distributions to be reconstructed. To overcome this problem, a symmetric encode–decoder fully convolutional network is proposed in this paper to improve the reconstruction quality of MART. The input of the neural network is the particle field reconstructed by the MART approach, while the output is the regenerated image with the same resolution. Numerical evaluations indicate that those blurred or irregular particles can be significantly refined by the trained neural network. Most of the ghost particles can also be removed by this filtering method. The reconstruction accuracy can be improved by more than 10% without increasing the computational cost. Experimental evaluations indicate that the trained neural network can also provide similar satisfactory reconstruction and improved velocity fields.
收稿日期: 2019-11-04 出版日期: 2020-04-03
通讯作者: Chao Xu     E-mail: cxu@zju.edu.cn
服务　　
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章　　
Jiaming Liang
Shengze Cai
Chao Xu
Jian Chu

引用本文:

Jiaming Liang, Shengze Cai, Chao Xu, Jian Chu. Filtering enhanced tomographic PIV reconstruction based on deep neural networks. IET Cyber-Systems and Robotics, 2020, 2(1): 43-52.

链接本文:

http://www.zjujournals.com/iet-csr/CN/10.1049/iet-csr.2019.0040    或    http://www.zjujournals.com/iet-csr/CN/Y2020/V2/I1/43

No related articles found!