学科云端学术报告(左超 教授,南京理工大学)  
 

发布者:吴琼   发布时间:2020-08-01  浏览次数:10    

【报告题目】基于学习的条纹投影轮廓术

【报  告 人】左超 教授 (南京理工大学)

【报告时间】85日(周三)1900开始

【直播地址】https://meeting.tencent.com/s/EFQiVsiP032Q

【会  议 ID729369976

【报告摘要】

Deep  learning is transforming a range of disciplines and eclipsing the state of the  art achieved by earlier machine-learning techniques. It has created new  opportunities to revolutionize optical metrology techniques. In this talk, we  introduce our recent efforts to apply deep-learning approaches to fringe  projection profilometry (FPP). We show that the deep-learning-enabled fringe  analysis approach can significantly boost the accuracy and improve the quality  of the phase reconstruction compared to conventional Fourier transform and  windowed Fourier transform approaches. Deep learning can also be used to perform  phase unwrapping and outperform conventional multi-frequency temporal phase  unwrapping in terms of both unwrapping reliability and robustness.As  a result, with the aid of deep learning, we can use less or even single raw  image for absolute phase retrieval, which enable FPP techniques to go a step  further in high-speed and high-accuracy 3D surface imaging of transient  events.

深度学习为光学测量技术的新一轮革新创造了新的机会。在这个报告中,我们介绍了我们最近将深度学习方法应用于条纹投影轮廓术的一系列工作。我们展示了与传统的傅里叶变换和加窗傅里叶变换法相比,深度学习的条纹分析方法可以显著提高相位重建的准确性和质量。深度学习还可以被用来进行时域相位解包裹,并且在解包可靠性和鲁棒性方面都优于传统的多频时域相位解包裹方法。在深度学习的帮助下,我们可以使用更少甚至单帧条纹图像实现绝对相位获取,这使得条纹投影轮廓术在高速、高精度甚至瞬态三维成像方面应用得以更进一步。

【报告人简介】

 Chao Zuo
 Nanjing University of Science and  Technology

Dr. Chao Zuo is a professor at the department of Electronic and  Optical Engineering, Nanjing University of Science and Technology (NJUST). He is  the principal investigator of the Smart Computational Imaging Laboratory  (SCILab: www.scilaboratory.com) at NJUST where the research interest focuses on  computational imaging and optical information processing. He has authored over  120 peer-reviewed journals publications, including more than 70 papers published  in JCR Q1 journals, e.g. Opt Lett, Opt Express, with over 4,000 citations. He is  Topical Editor of PhotoniX (10/2019-), Associate Editor of IEEE Access  (03/2019-), Microwave and Optical Technology Letters (03/2019-). He has been  selected into the Natural Science Foundation of China (NSFC) for Excellent Young  Scholars and Outstanding Youth Foundation of Jiangsu Province.  

     左超,南京理工大学电子工程与光电技术学院教授、博士生导师。南京理工大学智能计算成像实验室(SCILab:  www.scilaboratory.com)学术带头人与南京理工大学智能计算成像研究院(SCIRI:  www.njust-sci.com)执行董事(PI)。研究方向为计算光学成像与光信息处理技术。近年来已在SCI源刊上发表论文130余篇,包括JCR一区论文80余篇。6篇论文入选ESI高被引论文,2篇论文入选ESI热点论文,论文Google  Scholar引用逾4000次。研究成果12次被选作Adv. Photon., Opt.  Lett.Opt. Express等期刊的封面文章,并多次被SPIE NewsroomOSA Image of the  Week等报道。研究成果获国家发明专利57项,PCT国际专利11项,美国专利5项。国家“优秀青年科学基金”、江苏省“杰出青年基金”获得者。现任PhotoniX Topical Editor,  Microwave and Optical Technology LettersIEEE AccessAssociate  Editor,《光学学报》、《激光与光电子学报》专题编辑,《红外与激光工程》、中国激光杂志杂志社青年编委等。

 

[关闭窗口]