FPGA Resampling for Swept Source OCT Depth
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Swept Source OCT systems face challenges with high memory and processing power requirements for resampling algorithms, particularly for real-time processing, due to the need for Fast Fourier Transforms (FFTs), which leads to increased costs and insufficient performance from general-purpose processors.
Innovation Solution
A system utilizing a field programmable gate array (FPGA) to mathematically resample signals from a wide free spectral range reference interferometer, multiplying the reference clock rate to achieve greater imaging depth and enabling real-time performance with minimal sweep latency, while reducing harmonic distortion through direct convolution with finite impulse response digital filters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Fast Fourier Transforms (FFTs) are used for resampling in Swept Source OCT systems, then imaging depth and resolution are improved, but memory and processing power requirements increase significantly
Solution Approach 1:
The patent segments the resampling process into two distinct stages: first, a coarse resampling using a low-frequency K-Clock to capture the general signal structure, and second, a fine resampling using extracted instantaneous phase information to achieve the required precision. This segmentation allows the system to achieve high imaging depth and resolution without requiring excessive memory and processing power for a single large-scale FFT operation.
Solution Approach 2:
The patent replaces the traditional mechanical/computational FFT-based resampling system with a phase-extraction-based system. Instead of relying on high-frequency sampling and subsequent FFT processing that demands substantial computational resources, the system extracts instantaneous phase information from the low-frequency sampled signal and uses this phase data to reconstruct the high-resolution signal, thereby substituting a computationally intensive mechanism with a more efficient phase-based approach.
2Measurement precision
If a high frequency K-Clock is used to sample OCT signals, then sampling accuracy is improved, but hardware complexity and cost increase
Solution Approach 1:
The patent introduces an intermediary process between low-frequency sampling and high-resolution reconstruction: the extraction and unwrapping of instantaneous phase information. The low-frequency K-Clock samples the signal, and the intermediary phase extraction process recovers the high-frequency information embedded in the phase, eliminating the need for direct high-frequency hardware sampling while maintaining sampling accuracy.
Solution Approach 2:
The patent changes the parameter domain from time-frequency sampling to phase-domain representation. By extracting instantaneous phase information from the low-frequency sampled signal and using phase unwrapping techniques, the system transforms the problem from one requiring high-frequency temporal sampling to one that can be solved through phase parameter analysis, thereby reducing hardware complexity and cost.
3Productivity
If real-time processing is implemented with high processing power, then imaging speed is improved, but system cost and power consumption increase
Solution Approach 1:
The patent replaces the computationally intensive real-time FFT processing mechanism with a phase-based reconstruction approach. By extracting instantaneous phase information from low-frequency samples and using phase unwrapping followed by resampling with an interpolation factor, the system achieves real-time imaging performance without requiring high processing power, thereby reducing system cost and power consumption while maintaining imaging speed.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Real-time swept source OCT data is most often sampled using a specially cut hardware k- clock. The present invention involves mathematically resampling signals within an FPGA-based data acquisition board based on data sampled from a wide free spectral range reference interferometer. The FPGA can then multiply up the reference clock rate to achieve greater imaging depth. The Nyquist fold-over depth can thus be programmed from a standard reference to an arbitrary depth, much as PLL frequency synthesizer can produce many frequencies from a standard stable reference. The system is also capable of real-time performance.