Super Registration Method for Satellite Image Spectral Purity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods struggle to achieve precise alignment of satellite or aerial images, leading to geometric errors and 'jitter' when corrected to ground coordinates, which can be costly and difficult to resolve with high accuracy.
Innovation Solution
A method for registration of ortho corrected images involves obtaining and matching ortho images to a reference image, filtering out blunder match points, interpolating missing points, and resampling to generate a super registered image, which minimizes geometric errors and preserves spectral purity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If standard ortho correction techniques are used to remove geometric distortion, then geometric errors are reduced, but residual errors cause image jitter and misalignment when images are stacked
Solution Approach 1:
The patent implements an iterative feedback mechanism where match points are identified between images, residual misalignment is measured, and correction transformations are applied to reduce the misalignment. This feedback loop continues until alignment converges to the desired precision level, resolving the residual jitter problem while maintaining geometric accuracy.
Solution Approach 2:
The patent applies dynamic correction transformations that adapt to local variations in misalignment across different regions of the image. Rather than using a static global transformation, the system dynamically adjusts corrections based on locally identified match points, enabling sub-pixel alignment precision while preserving the geometric integrity established by ortho correction.
2Manufacturing precision
If resampling is applied during ortho correction to align images to ground coordinates, then geometric alignment is improved, but spectral purity of the imagery is degraded
Solution Approach 1:
The patent applies resampling only partially - specifically for the final sub-pixel alignment adjustment after ortho correction, rather than during the entire ortho correction process. This limited application of resampling achieves the necessary alignment precision while minimizing spectral degradation by avoiding repeated resampling operations.
Solution Approach 2:
The patent performs preliminary ortho correction using integer-pixel transformations before applying the final sub-pixel alignment. By completing the major geometric correction first with methods that preserve spectral purity, and then applying minimal resampling only for fine alignment, the system achieves both precision and spectral preservation.
3Measurement precision
If highly accurate geometric data is obtained to achieve sub-pixel alignment, then image jitter is reduced, but cost and difficulty of data acquisition increase
Solution Approach 1:
The patent enables the system to self-correct alignment errors by automatically identifying match points between images and computing correction transformations from the images themselves. This self-service approach eliminates the need for expensive external geometric data sources, achieving sub-pixel alignment using only the image data and automated feature matching.
Solution Approach 2:
The patent replaces the mechanical/surveying-based system for obtaining accurate geometric data (such as ground control points and precise orbital information) with an automated image-processing system that uses feature matching and computational algorithms to achieve the same alignment precision at lower cost and without specialized equipment.
Data Source
AI summary
Provided are systems and methods for registration and carrying out fine image adjustments of aerial or satellite images to obtain super registered images. Also provided are systems and methods for registration where resampling is minimized or avoided to reduce, minimize, or prevent spectral degradation.