Smart Computational Imaging (SCI) Lab
智能计算成像实验室

天气信息
文章分类

文章分类

副标题

会员登录
登录
研究亮点

Thesynergy of traditional techniques and deep learning enables single‐framehighprecision fringe pattern analysisA new publication from Opto Electronic Advances, 10.29026/oea.2024.230034 discusses h...

Recently, deep learning has yielded transformative success across optics and photonics, especially in optical metrology. Deep neural networks (DNNs) with a fully convolutional architecture (e.g., U-Ne

Recently, deep learning has yielded transformative success across optics and photonics, especially in optical metrology. Deep neural networks (DNNs) with a fully convolutional architecture (e.g., U-Ne

La métrologie optique, en tant que technique de métrologie à usage général utilisant la lumière comme support d'informations pour des mesures sans contact et non destructives, est fondamentale pour le

Optical metrology, as a general-purpose metrology technique that uses light as information carriers for non-contact and non-destructive measurement, is fundamental to manufacturing, basic research,...

南京理工大学陈钱、左超教授课题组提出了一种结合传统技术与深度学习的单帧高精度条纹图案分析方法。通过将传统单帧条纹分析方法嵌入到深度学习的网络结构和损失函数中,利用可学习的自适应频域滤波直接从单幅条纹图像中恢复出可靠的相位初值,进而引导神经网络实现高精度、高计算效率的相位测量。这项研究表明,基于物理模型的传统技术和数据驱动的深度学习方法协同作用有望为实现快速、高精度的单帧三维成像开辟新途径。

SCILab 官方公众号
SCILab 官方B站