Target-less Wheel Alignment Image Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Machine vision vehicle wheel alignment systems without predefined alignment targets face challenges in accurately determining vehicle wheel alignment angles due to the variability of features observed in images from vehicle to vehicle and wheel assembly to wheel assembly, requiring robust image processing methods to accommodate these differences.
Innovation Solution
The method involves identifying corresponding image features on a vehicle wheel assembly in multiple images, using cylindrical symmetry optimization and superquadric or hyperquadric equations to fit mathematical models to the observed features, allowing for the determination of the wheel assembly's pose and alignment angles without the need for predefined targets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If target-less machine vision systems are used to observe wheel assemblies, then the system can accommodate variability between vehicles and wheel assemblies, but the measurement precision deteriorates due to the randomness of observable features
Solution Approach 1:
The patent transforms the random 2D image features into structured 3D spatial parameters by fitting superquadric mathematical models to the wheel assembly features. This parameter transformation converts variable appearance features into consistent geometric parameters (center position, orientation angles, dimensions) that enable precise measurement despite feature variability across different vehicles and wheel assemblies.
Solution Approach 2:
The patent introduces mathematical models (superquadric equations) as intermediaries between the observed image features and the alignment measurements. These models serve as a bridge that translates variable visual features into standardized geometric parameters, enabling consistent measurement precision while maintaining adaptability to different wheel assembly configurations.
2Ease of operation
If alignment targets are not mounted on wheels, then the system is easier to operate and requires no additional components, but the reliability of feature identification deteriorates due to feature variability
Solution Approach 1:
The patent enables the wheel assembly to serve itself as the alignment target by automatically identifying and utilizing its own geometric features (wheel rim, tire sidewalls, hub structures) visible in the images. The system processes the wheel assembly's inherent features through mathematical modeling to determine alignment parameters, eliminating the need for external target mounting while maintaining reliable feature identification through robust image processing algorithms.
3Measurement precision
If traditional alignment targets are used, then measurement precision is maintained through predefined features, but the device complexity increases due to target mounting and removal procedures
Solution Approach 1:
The patent extracts the essential measurement function from the physical alignment target and embeds it directly into the image processing algorithm. By removing the external target component and incorporating the feature identification and measurement capabilities into the software processing pipeline, the system maintains measurement precision while eliminating the mechanical complexity of target mounting, attachment, and removal mechanisms.
Data Source
AI summary
A method of the present invention provides methods for processing images of vehicle wheel assemblies acquired during vehicle service procedures to determine vehicle wheel assembly pose components, from which vehicle wheel alignment angles may be calculated.