Polynomial Camera Model Calibration Using Passive Targets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing camera calibration methods for ray-based camera models are complex and often require active targets with perfect geometry, or they allow only insufficient calibration, particularly for central and non-central imaging systems.
Innovation Solution
An apparatus and method that derive a camera model by obtaining a calibration image, creating a sparse image with image points, mapping these points to ray support points using a polynomial function, and generating a camera model based on these points, allowing for accurate calibration of both central and non-central camera systems using passive targets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If model based camera calibration is used, then the calibration process is simple, but the accuracy is insufficient because the parametric model is only approximately true
Solution Approach 1:
The patent introduces a passive target with known geometry as an intermediary object between the camera and the calibration process. This target serves as a mediator that provides precise reference points for ray mapping without requiring complex active targets or assumptions about camera projection models. The target's known geometry acts as an intermediate reference frame that enables accurate calibration while keeping the process simple.
2Adaptability or versatility
If existing ray based camera calibration methods are used, then no assumption about projection type is made (black box approach), but the calibration is either complex requiring active targets with perfect geometry or allows only insufficient calibration
Solution Approach 1:
The patent replaces expensive and complex active targets with perfect geometry with a simple passive target that has known geometry. This passive target is essentially a 'cheap' alternative that doesn't require active components, complex mechanisms, or perfect manufacturing tolerances. The target can be a simple printed pattern or geometric structure that provides sufficient reference points for calibration without the complexity of active targets.
3Adaptability or versatility
If existing ray based camera calibration methods are used, then projection model independence is achieved, but calibration accuracy is insufficient
Solution Approach 1:
The patent changes the fundamental parameter approach by using a polynomial function to describe the mapping between image coordinates and target coordinates, rather than relying on traditional parametric camera models. This polynomial parameterization allows for flexible, accurate representation of the projection without assuming a specific camera model, thereby achieving both projection independence and high accuracy through the flexible polynomial parameters.
Data Source
AI summary
The disclosure pertains to an apparatus comprising a circuitry. The circuitry is configured to obtain a calibration image of a target, to derive a sparse image based on the calibration image, wherein the sparse image includes image points, to derive ray support points based on the image points by performing an image to target mapping of the image points based on a polynomial function, wherein the ray support points being indicative of light rays reflected by the target and incidenting on an image sensor, and to generate a camera model based on the derived ray support points.