Compressive Imaging Beyond One Trillion Frames Per Second

每秒超过一万亿帧的压缩成像

• 批准号: 1609693
• 负责人: Mark Foster
• 金额: $ 28万
• 依托单位: Johns Hopkins University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2016
• 资助国家: 美国
• 起止时间: 2016-07-01 至 2019-12-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1609693&HistoricalAwards=false
• 关键词: Compressive; Imaging; Beyond; Trillion; Frames

This research program will develop a video imaging system operating at frame rates beyond one trillion frames per second and that is capable of recording isolated (non-repetitive) events. No current camera technology can operate at these extremely high frame rates and video durations for the observation of isolated events. The primary challenges for such a technology are the short time-gating and high light irradiance necessary to achieve fast exposure times with extremely high frame rate and the massive information bandwidth associated with such high-speed image acquisition. Such a high-speed single-shot imaging system is an enabling technology for numerous applications throughout engineering and the physical and life sciences. In particular, we plan to leverage the ultrafast single-shot imager developed through this program to better understand the dynamics of materials under extreme conditions. This research can positively impact society from a health and safety perspective through better understanding of the effect of impacts on materials and thus facilitate the development of materials that can better control and prevent injuries to the human body. In addition, through this research program we plan to offer undergraduate research experiences, provide the project data and experimental system for educational purposes such as class design projects and hands-on lab demonstrations, and further our participation in outreach activities for young students from severely underrepresented groups within the engineering discipline. The primary goal of this research program is to develop and experimentally validate a photonic imaging system employing compressed sensing (CS) for video acquisition at frame rates well beyond a terahertz (THz). Our approach leverages the dimensionality reduction afforded by CS to reconstruct three-dimensional spatio-temporal video information from a single high resolution two-dimensional image captured by a camera. Thereby, the measurement efficiency of CS will also mitigate the traditional challenge of the acquisition of a large amount of image data in an extremely short amount of time. Furthermore, our approach is built upon a temporal Fourier processor using a time-lens to imprint the temporal scene dynamics onto an ultrafast laser pulse's spectrum allowing for capture of three-dimensional spatio-temporal video information using a hyperspectral CS camera architecture. This time-lens approach fully leverages the available optical bandwidth maximizing both the frame rate and the number of frames captured in a single exposure. We aim to reach frame rates beyond 1 THz and record lengths of more than 100 frames. No current technology can achieve this combination of performance, yet such a technology can be revolutionary for understanding of ultrafast physical phenomena. Specifically, such single-shot imaging is necessary for observing isolated events such as destructive, rare, and/or costly events and is extremely challenging on ultrafast time-scales. Through this research program we will develop this imager, investigate methods for increasing image contrast, and begin to explore its application to the understanding of material failure under extreme conditions.

该研究计划将开发一种视频成像系统，其帧速率超过每秒一万亿帧，并且能够记录孤立（非重复）事件。目前没有摄像机技术可以在这些极高的帧速率和视频持续时间下操作，以观察孤立的事件。这种技术的主要挑战是短时间选通和高光辐照度，这是实现具有极高帧率的快速曝光时间所必需的，以及与这种高速图像采集相关的大量信息带宽。这种高速单次拍摄成像系统是整个工程、物理和生命科学中众多应用的一项使能技术。特别是，我们计划利用通过该计划开发的超快单次成像仪，以更好地了解极端条件下材料的动态。这项研究可以从健康和安全的角度对社会产生积极的影响，通过更好地了解影响对材料的影响，从而促进可以更好地控制和防止对人体伤害的材料的开发。此外，通过这项研究计划，我们计划提供本科生的研究经验，提供项目数据和实验系统的教育目的，如课堂设计项目和动手实验室演示，并进一步参与外展活动，为年轻学生从严重代表性不足的群体在工程学科。该研究计划的主要目标是开发并实验验证一种采用压缩传感（CS）的光子成像系统，以远高于太赫兹（THz）的帧速率进行视频采集。我们的方法利用CS提供的降维重建三维时空视频信息从一个单一的高分辨率的二维图像由摄像机捕获。因此，CS的测量效率也将减轻在极短时间内采集大量图像数据的传统挑战。此外，我们的方法是建立在一个时间傅立叶处理器使用时间透镜的时间场景的动态压印到超快激光脉冲的光谱，允许使用高光谱CS相机架构的三维时空视频信息的捕获。这种时间透镜方法充分利用了可用的光学带宽，最大限度地提高了帧速率和单次曝光中捕获的帧数量。我们的目标是达到超过1 THz的帧速率和超过100帧的记录长度。目前没有任何技术可以实现这种性能组合，但这种技术对于理解超快物理现象来说可能是革命性的。具体而言，这种单次拍摄成像对于观察孤立事件（诸如破坏性事件、罕见事件和/或昂贵事件）是必要的，并且在超快时间尺度上是极其具有挑战性的。通过这项研究计划，我们将开发这种成像仪，研究增加图像对比度的方法，并开始探索其在极端条件下材料失效的理解中的应用。