Camera Calibration Using Normal Vectors and Virtual Planes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing camera calibration methods face accuracy issues when obstacles are present in the real space, requiring large calibration equipment and complex 3D coordinate measurements, especially in wide and complicated target areas.
Innovation Solution
A camera calibration device that estimates camera parameters using freely-arranged reference points, acquiring normal vectors from images to estimate rotation and translation matrices without the need for special calibration equipment or 3D coordinate measurements, by projecting vectors on a virtual plane and optimizing based on perpendicularity and re-projection errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If special calibration equipment with known arrangement of reference points is installed, then camera parameters can be estimated using projection equations, but installation place is restricted by obstacles and reference points cannot be evenly acquired from shot image
Solution Approach 1:
The patent uses markers (copies of reference points) placed on existing objects in the environment instead of requiring special calibration equipment. These markers can be freely arranged on any objects within the camera's field of view, eliminating the need for dedicated calibration equipment and allowing calibration in environments with obstacles.
Solution Approach 2:
The patent separates the reference points from their traditional calibration equipment context by placing markers on various existing objects in the environment. This segmentation allows reference points to be distributed throughout the entire field of view rather than being confined to a single calibration device, enabling even distribution of reference points across the image.
2Measurement precision
If piece of calibration equipment is used, then reference points arrangement is known, but when target real space is wide, practically-impossible large calibration equipment is required to acquire reference points from everywhere on the screen
Solution Approach 1:
The patent makes the calibration system universal by allowing markers to be placed on any objects within the environment rather than requiring a specific calibration device. This multi-functionality enables calibration in both small and large spaces using the same marker-based approach, eliminating the need for oversized calibration equipment in wide target spaces.
Solution Approach 2:
The patent transitions from using a single calibration device in one location to distributing multiple markers throughout the three-dimensional space being calibrated. By placing markers on various objects at different positions and orientations, the system achieves comprehensive coverage of wide target spaces without requiring excessively large calibration equipment.
3Ease of operation
If freely-arranged reference points are used, then easiness of camera calibration is improved, but 3D coordinates have to be measured one by one in the real space which is very complicated in wide and complicated target space
Solution Approach 1:
The patent uses markers as simplified copies of reference points that can be easily identified in images. These markers have known geometric characteristics that allow automatic detection and 3D coordinate calculation without manual measurement, maintaining ease of operation while eliminating the complexity of measuring 3D coordinates for freely-arranged reference points.
Solution Approach 2:
The patent changes the parameters of the reference points by using markers with specific, known geometric properties (such as circular shapes or specific patterns). This allows the system to automatically determine 3D coordinates from 2D image data through geometric constraints, eliminating the need for manual 3D measurement while preserving the flexibility of freely-arranged reference points.
Data Source
Figure 1
Figure 2(a)~2(b)
Figure 3
AI summary
There is provided a camera calibration device capable of highly precisely and easily estimating camera parameters without installing a piece of special calibration equipment or measuring 3D coordinates of reference points employed for calibration. A normal vector acquisition means 81 acquires normal vectors perpendicular to a reference horizontal plane from an image of a camera to be calibrated. A rotation matrix estimation means 82 projects the acquired normal vectors on a projection virtual plane perpendicular to the reference horizontal plane and evaluates that the projected normal vectors are perpendicular to the reference horizontal plane thereby to estimate a rotation matrix employed for calibrating the camera.