Automated Image Registration via Radiometric and Geometric Tie Point Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Manual image registration is laborious, prone to errors, and becomes prohibitive for large datasets due to the complexity of remote sensing images, which are affected by navigation errors, atmospheric scattering, and varying image characteristics such as multi-temporal effects and terrain distortions.
Innovation Solution
An automated image registration method using a registration engine with components like map information, manual and automatic tie point generation, sensor models, and epipolar geometry constraints, along with template matching and filtering techniques to improve accuracy and reliability, allowing for batch processing and integration with other automated systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual registration is used to locate and match feature points between images, then registration accuracy can be maintained, but the process becomes laborious, tedious and prohibitive for large datasets
Solution Approach 1:
The system performs automatic tie point generation and image registration without human intervention. The computer executes algorithms that automatically locate and match feature points between images, eliminating the need for manual operator input while maintaining registration accuracy through automated feature detection and matching processes
Solution Approach 2:
The patent replaces the manual mechanical process of visual inspection and point matching with automated computational algorithms. The system uses computer-based image processing, coordinate transformation, and feature matching algorithms to substitute human operators, dramatically improving processing efficiency while maintaining or enhancing registration precision
2Productivity
If automatic tie point generation is attempted, then processing efficiency improves, but location errors occur due to navigation errors, atmospheric scattering, and image artifacts
Solution Approach 1:
The system incorporates feedback mechanisms where the registration engine continuously evaluates the quality and reliability of generated tie points. Based on feedback from image quality assessment, navigation data validation, and matching confidence scores, the system can iteratively refine tie point selection, reject unreliable matches, and adjust processing parameters to maintain accuracy while preserving automated processing efficiency
Solution Approach 2:
The patent dynamically adjusts processing parameters such as feature detection thresholds, matching confidence levels, and search window sizes based on image characteristics and quality metrics. By changing these parameters adaptively, the system optimizes the balance between processing speed and tie point accuracy, maintaining reliability across varying image conditions while preserving automated efficiency
3Measurement precision
If comprehensive feature matching is performed across the entire image, then registration accuracy improves, but the search space becomes too large for practical processing
Solution Approach 1:
The patent divides the image into multiple regions or zones, and performs feature matching independently within each segment. This segmentation approach reduces the computational search space from the entire image to smaller local regions, maintaining matching accuracy within each segment while dramatically reducing overall computational complexity and enabling practical processing of large images
Data Source
AI summary
A computer-implemented method and system register plural images using a computer processor and computer memory, where computer code stored in the computer memory causes the computer processor to perform the registration. The registration includes receiving a reference image; receiving a sensed image; computing an approximate image transformation between the reference image and the sensed image; finding plural candidate tie points by utilizing template matching; applying radiometric filtering to at least a portion of the plural candidate tie points; applying geometric filtering to at least a portion of the plural candidate tie points; computing a calculated image transformation using candidate points of the plural candidate points that were not filtered out by the radiometric and geometric filtering; transforming the sensed image using the calculated image transformation; and registering the sensed image with the reference image.