Real-time Super Resolution for Long Standoff Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional wide area motion imagery systems face challenges in achieving high spatial resolution and desired signal-to-noise ratios at long standoff ranges due to the need for large optical systems, which often result in performance trade-offs between field-of-regard, revisit time, and cost considerations.
Innovation Solution
A system that combines a focal plane array with a fast steering mirror and a real-time super resolution module to estimate shifts, rotations, and zooms between images, generate a common super resolution image frame, and mitigate the impact of bad pixels, using parallel processing to produce high-resolution images at long standoff distances.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a large optical system is used to achieve high spatial resolution and good signal-to-noise ratio at long standoff ranges, then measurement precision and signal-to-noise ratio are improved, but device complexity and cost increase
Solution Approach 1:
The patent segments the imaging task into multiple lower-resolution frames captured at different positions, then computationally reconstructs a high-resolution image from these segments. This avoids the need for a single large optical system by distributing the resolution achievement across multiple smaller captures.
Solution Approach 2:
The patent replaces the mechanical/optical solution (large aperture optics) with a computational solution (super-resolution algorithms). Instead of relying on physical optics to achieve high resolution, the system uses software processing to synthesize high-resolution images from multiple lower-resolution inputs.
2Reliability
If a large optical system is used to achieve good signal-to-noise ratio at long standoff ranges, then signal-to-noise ratio is improved, but device complexity and cost increase
Solution Approach 1:
The patent merges multiple noisy frames captured at different positions into a single high-resolution image with improved signal-to-noise ratio. By combining information from multiple captures, the system achieves better reliability without requiring a large optical system.
Solution Approach 2:
The patent replaces the mechanical/optical solution (large aperture for light gathering) with a computational solution (frame integration algorithms). Instead of relying on physical optics to improve signal-to-noise ratio, the system uses software to combine multiple frames and reduce noise.
3Adaptability or versatility
If the field-of-regard is increased to cover a large area, then adaptability is improved, but measurement precision deteriorates
Solution Approach 1:
The patent segments the large field-of-regard into multiple smaller regions captured at different positions, then reconstructs high-resolution images for each region. This allows the system to maintain both wide coverage and high resolution by processing different areas separately.
Solution Approach 2:
The patent adds the temporal dimension to the imaging process by capturing frames over time at different positions. This allows the system to achieve high resolution in a small region of interest while maintaining a large overall field-of-regard, as different regions are imaged at different times.
4Productivity
If the revisit time is decreased to achieve faster frame rates, then productivity is improved, but measurement precision deteriorates
Solution Approach 1:
The patent segments the high-resolution imaging task into multiple lower-resolution frames captured rapidly in sequence. By capturing multiple frames quickly and then reconstructing them computationally, the system achieves both fast revisit times and high final resolution.
Solution Approach 2:
The patent performs preliminary captures of multiple frames at lower resolution and faster rates, then processes them computationally to produce the final high-resolution image. This preliminary action allows the system to gather necessary data quickly before performing the computationally intensive reconstruction.
Data Source
AI summary
The system and method for super resolution processing at long standoff distances in real-time. The system collects a series of image frames and estimated the shift, rotation, and zoom parameters between each of the image frames. A matrix is generated and then an inversion is applied to the matrix to produce a super resolution image of an area of interest while mitigating the effect of any bad pixels on image quality. In some cases, the area of interest is user-defined and in some cases image chips are provided by tracking software. A fast steering mirror can be used to steer and/or dither the focal plane array.