Distance Measurement Camera Calibration for Windshield Distortion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In-vehicle distance measurement cameras installed behind windshields face challenges due to windshield distortion, requiring calibration that is inaccurate when performed with the camera mounted in a vehicle, and large calibration objects are needed for long-distance calibration, increasing device size.
Innovation Solution
A calibration method for distance measurement devices that captures images with and without a windshield, using disparity between images from multiple viewpoints to calculate correction information, and modifies this information based on images captured through the windshield to account for distortion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If calibration is performed with the camera mounted in the vehicle through the windshield, then the calibration accounts for windshield distortion, but the calibration accuracy is insufficient due to the complex curved surface shape of the windshield
Solution Approach 1:
The calibration process is divided into two separate stages: first calibrating the camera without the windshield to obtain baseline parameters, then calibrating through the windshield to capture distortion characteristics. This segmentation allows each calibration stage to focus on specific aspects, improving overall accuracy despite the complex windshield geometry.
Solution Approach 2:
The camera is pre-calibrated without the windshield before being mounted in the vehicle. This preliminary calibration establishes accurate baseline parameters that are then used as reference during the second calibration stage through the windshield, compensating for the distortion introduced by the curved surface.
2Adaptability or versatility
If the calibration object is positioned at a long distance to cover the entire image angle, then the calibration range is improved, but the size of the calibration object must be increased
Solution Approach 1:
The calibration method transitions from relying solely on the physical size and distance of the calibration object to utilizing the optical properties of the windshield itself as a calibration reference. By treating the windshield's distortion characteristics as the calibration dimension, the system achieves long-distance calibration capability without requiring proportionally larger calibration objects.
Solution Approach 2:
The windshield acts as an intermediary element in the calibration process. Instead of directly calibrating the camera with distant objects, the system uses the windshield's known distortion characteristics as an intermediate reference, allowing accurate calibration across the entire image angle without requiring excessively large calibration objects at long distances.
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 allows for accurate calibration of in-vehicle distance measurement cameras without increasing device size, by separately calculating and modifying correction values to compensate for windshield distortion, enhancing distance measurement precision.
Implementation Method 1
calculates a distance to an object by using disparity (parallax) between images captured from at least two points of view
Data Source
AI summary
A calibration method for a distance measurement device that is mounted inside a moving body, images outside of the moving body without intervention of a transparent body, and calculates a distance to an object by using disparity between images captured from at least two points of view, the method including: a first process of capturing a first-image-for-calibration and a second-image-for-calibration at different distances between a first-object-for-calibration and the distance measurement device without intervention of the transparent body and calculating correction information for converting disparity information calculated from image information of each of the first-image-for-calibration and the second-image-for-calibration into distance information; and a second process of capturing a third-image-for-calibration from a second-object-for-calibration that is located at least at one distance via the transparent body and modifying the correction information calculated in the first process on the basis of image information of at least the captured third-image-for-calibration.