Camera Calibration via View Angle Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional camera calibration methods struggle to achieve accurate calibration across the entire field of view, particularly for wide angle cameras, due to increased calibration errors at the edge of the field of view, leading to reduced distance measurement accuracy in stereo distance measurement.
Innovation Solution
A camera calibration method that calculates camera parameters using a calibration point, involving the acquisition of three-dimensional coordinate sets and image coordinate pairs, calculation of view angle-corresponding lengths, three-dimensional positions, weighting of measurement points, and updating of camera parameters to achieve accurate calibration across a large area.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional camera calibration methods are used, then calibration can be performed, but calibration accuracy deteriorates at the edge of the field of view for wide angle cameras
Solution Approach 1:
The patent applies local quality by weighting calibration points differently based on their position in the field of view. Calibration points at the edge of the field of view are given different weights compared to those at the center, allowing the calibration process to account for the varying accuracy characteristics across different regions of the image sensor.
Solution Approach 2:
The patent changes the parameter of weight assignment for calibration points based on their spatial position. By introducing position-dependent weighting factors, the calibration algorithm adapts to the non-uniform error distribution across the field of view, improving overall calibration accuracy particularly at the edges.
2Area of stationary object
If wide angle cameras are used to increase field of view coverage, then area coverage is improved, but distance measurement accuracy deteriorates due to calibration errors
Solution Approach 1:
The patent addresses this contradiction by applying local quality through position-dependent weighting. Wide angle cameras capture a large field of view, but calibration accuracy varies across the image. The method weights calibration points according to their position, giving appropriate emphasis to edge points to compensate for the increased distortion and calibration difficulty inherent in wide angle optics.
Solution Approach 2:
The patent replaces the conventional uniform calibration approach with a computational weighting mechanism. Instead of relying solely on mechanical precision or uniform mathematical treatment of all calibration points, the system uses algorithmic weight assignment based on spatial position to correct for the inherent limitations of wide angle geometry.
3Area of stationary object
If calibration points at the edge of the field of view are used, then field of view coverage is improved, but calibration error increases
Solution Approach 1:
The patent changes the parameter of weight assignment for calibration points based on their spatial position. By introducing position-dependent weighting factors, the calibration algorithm adapts to the non-uniform error distribution across the field of view, improving overall calibration accuracy particularly at the edges.
Solution Approach 2:
The patent converts the harmful effect of increased calibration error at edge points into a benefit by deliberately weighting these points in the calibration process. Rather than ignoring edge points due to their higher error potential, the method incorporates them with appropriate weighting, thereby utilizing all available calibration data to improve the overall calibration model.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This method enables precise calibration of multiple-lens cameras, reducing distance measurement errors and improving accuracy across the entire field of view by weighting measurement points based on view angle-corresponding lengths, thereby minimizing the influence of calibration errors.
Implementation Method 1
calculating a three-dimensional position of a measurement point for each pair of the cameras by use of parallax of the calibration point between the cameras in the pair of the cameras
Data Source
Figure 1
Figure 2~3
Figure 4A~4B
AI summary
A camera calibration method which calculates camera parameters of at least three cameras acquires three-dimensional coordinate set of a calibration point and image coordinate pair of the calibration point in each camera image, acquires camera parameters of each camera, calculates a view angle-corresponding length (L1b, L1a) corresponding to a view angle of each pair of cameras (21, 22) viewing the calibration point (P1), calculates a three-dimensional position of a measurement point corresponding to a three-dimensional position of the calibration point for each camera pair (21,22) using parallax of the calibration point between the cameras in the camera pair (21,22), weights the three-dimensional position of each measurement point using the view angle-corresponding length (L1b, L1a) corresponding to the measurement point (P1), calculates a three-dimensional position of a unified point of the weighted measurement points, and updates the camera parameters based on the three-dimensional coordinate set of the calibration point and the three-dimensional position of the unified point.