Rolling-Shutter Sensor Geometric Distortion Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional space-based imaging systems using rolling-shutter framing sensors face complex geometric distortions due to platform motion, which existing methods struggle to accurately correct, especially in high-velocity applications, leading to inefficiencies and increased size, weight, and cost of satellite platforms.
Innovation Solution
The development of a method and system that generates a DtoN mapping to transform distorted image space to a normalized image space, allowing for the identification of pixel positions and subsequent interpolation and resampling to produce a corrected image, utilizing a digital signal processor to manage the mapping and resampling process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If rolling-shutter framing sensors are used to reduce cost and complexity, then device complexity and cost are reduced, but geometric distortion correction becomes significantly more difficult and inaccurate
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing a lookup table (LUT) that maps distorted pixel coordinates to corrected coordinates before image processing. This pre-computed mapping table is generated based on the known rolling-shutter distortion model and platform motion parameters, allowing rapid correction during actual imaging operations without real-time complex calculations.
Solution Approach 2:
The patent introduces an intermediary lookup table that acts as a mediator between the distorted rolling-shutter image and the corrected output. This LUT contains pre-computed transformation data that simplifies the complex coordinate transformation problem into a straightforward table lookup and interpolation operation, making the distortion correction process computationally efficient.
2Weight of stationary object
If rolling-shutter framing sensors are used instead of global-shutter sensors, then size, weight, and cost of the platform are reduced, but geometric distortion increases
Solution Approach 1:
The patent replaces complex mechanical solutions (such as using heavier global-shutter sensors or more sophisticated platform stabilization mechanisms) with a computational approach. By using software-based distortion correction algorithms and pre-computed lookup tables, the system achieves geometric accuracy without requiring additional mechanical components or heavier hardware.
3Manufacturing precision
If complex distortion correction algorithms are implemented in real-time, then image accuracy is improved, but processing time and computational resources increase
Solution Approach 1:
The patent pre-computes the distortion correction lookup table based on the rolling-shutter distortion model and expected platform motion parameters. This allows the actual image correction to be performed through simple table lookup and interpolation operations rather than complex real-time calculations, significantly reducing processing time while maintaining accuracy.
Solution Approach 2:
The patent uses a practical approximation approach by implementing a lookup table with sufficient resolution to achieve the required accuracy level. Rather than computing exact corrections for every possible pixel position, the system uses interpolated values from a coarser grid, providing adequate accuracy with reduced computational overhead.
Data Source
AI summary
Aspects and embodiments are generally directed to an imaging system and methods for correcting geometric distortion induced by rolling-shutter operation of framing sensors. In one example, a method includes the acts of developing a DtoN mapping from a distorted image space to a normalized image space for initial pixels of the rolling-shutter framing sensor, developing an NtoD mapping from the normalized image space to the distorted image space for repositioned pixels in the normalized image space based on the DtoN mapping, the repositioned pixels corresponding to the initial pixels of the rolling-shutter framing sensor, producing a normalized image based on the repositioned pixels and the NtoD mapping, and resampling the normalized image to produce a corrected image of the imagery collected by the rolling-shutter framing sensor.