Real-time quantitative measurement of OCT data using GPGPU

使用GPGPU实时定量测量OCT数据

• 批准号: 485511-2015
• 负责人: Sarunic, Marinko
• 金额: $ 1.82万
• 依托单位: Simon Fraser University
• 依托单位国家: 加拿大
• 项目类别: Engage Grants Program
• 财政年份: 2015
• 资助国家: 加拿大
• 起止时间: 2015-01-01 至 2016-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=585012
• 关键词: Real; time; quantitative; measurement; OCT

Fourier Domain Optical Coherence Tomography (FD OCT) is a micrometer-scale, cross-sectional imaging technology that is commonly described as an optical analogue to ultrasound imaging. FD OCT is growing rapidly into applications of industrial imaging and non-destructive testing (NDT). The computational load for FD OCT signal processing and image analysis is substantial. The FD OCT signal is acquired as interferometric data, which is converted to images via Fourier transform. Currently, measurements on the images are performed off line, limiting the applications of the technology. The purpose of this research proposal is to investigate a heterogeneous computing system combining FPGA and GPGPU for high speed volumetric FD OCT image acquisition and real-time analysis for NDT. We propose to use FPGA hardware for the FD OCT signal processing steps (converting interferograms to image data), followed by GPGPU for image segmentation, extraction of quantitative measurements, and display of results. This processing flow matches the repetitive routines to the FPGA platform, and the more complicated image segmentation algorithms and the display functions to the GPGPU. The project plan described in detail below will foster the establishment of a new collaboration, and facilitate the investigation into novel heterogeneous computing for real-time quantitative volumetric FD OCT.

傅里叶域光学相干层析成像（FDOCT）是一种微米级的截面成像技术 该技术通常被描述为超声成像的光学模拟。FD OCT正在增长 快速应用于工业成像和无损检测（NDT）。计算负荷为 FD OCT的信号处理和图像分析是实质性的。FD OCT信号作为干涉测量信号被采集。 数据，通过傅立叶变换转换为图像。目前，对图像的测量是 离线执行，限制了该技术的应用。 本研究的目的是探讨一种结合FPGA的异构计算系统 和GPGPU，用于高速体积FD OCT图像采集和无损检测实时分析。我们提出 为了将FPGA硬件用于FD OCT信号处理步骤（将干涉图转换为图像数据）， 然后是GPGPU，用于图像分割、定量测量的提取和结果显示。 这种处理流程将重复的例程匹配到FPGA平台上， 分割算法和显示功能的GPGPU。项目计划详述如下 将促进建立一个新的合作，并促进调查新的异质 实时定量体积FD OCT的计算。