Multi-Camera Imager Registration via Fiducial Grid Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Multi-camera imaging systems face challenges in accurately registering images from different component cameras, which is complex and often requires controlled factory conditions for calibration, making real-time field registrations difficult.
Innovation Solution
The system employs a simplified method for determining calibration and run-time registrations by using a calibration target with a planar rectilinear grid of fiducial icons, adjusting for parallax along epipolar lines, and applying polynomial mapping functions to generate a fundamental matrix for camera coordinate transformations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional factory calibration methods are used, then manufacturing precision is improved, but device complexity and operational flexibility deteriorate
Solution Approach 1:
The patent applies preliminary action by pre-calibrating cameras at the factory with controlled imaging conditions to establish initial registration parameters. These pre-determined parameters are then used during field operation to quickly register images without requiring complex real-time calibration procedures, thus maintaining manufacturing precision while simplifying operational complexity
Solution Approach 2:
The system employs self-service by using the cameras themselves to capture calibration images of a calibration target, which automatically provides the data needed for registration. The calibration target with known geometry allows the system to self-calibrate without external intervention, reducing operational complexity while maintaining accuracy
2Measurement precision
If controlled factory calibration conditions are used, then measurement precision is improved, but adaptability to field conditions deteriorates
Solution Approach 1:
The patent uses preliminary calibration actions performed under controlled factory conditions to establish baseline registration parameters. These pre-determined parameters serve as a foundation that can be efficiently adjusted for field conditions, enabling the system to adapt to various operational environments while maintaining measurement precision through the pre-established calibration framework
Solution Approach 2:
The system implements dynamics by allowing the calibration registration to be adjusted for parallax effects during field operation. The patent enables dynamic adaptation to different imaging conditions while maintaining the foundational accuracy established during factory calibration, thus achieving both precision and adaptability
3Manufacturing precision
If complex registration procedures are used, then manufacturing precision is improved, but ease of operation deteriorates
Solution Approach 1:
The patent applies preliminary action by performing complex calibration procedures during factory manufacturing to determine accurate camera coordinate transformations. These pre-computed registration parameters are then stored and applied during field operation, eliminating the need for operators to perform complex registration procedures manually, thus improving ease of operation while maintaining manufacturing precision
Solution Approach 2:
The system extracts the complex registration computation from the field operation phase and relocates it to the factory calibration phase. By taking out the computationally intensive tasks during manufacturing, the patent leaves simple, straightforward operations for field use, significantly improving ease of operation while preserving the accuracy achieved through complex initial calibration
Data Source
AI summary
A method of registering first and second cameras in a multi-camera imager comprising imaging a calibration target having rows and columns of fiducials with both the first and second cameras and using convergence angles of images of the rows and columns in the images to register the cameras.