Online Camera Calibration for Autonomous Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional camera calibration methods for autonomous driving vehicles are time-consuming and require manual intervention, often taking up to 30 minutes to complete, due to the need for offline calibration using a mobile shelf with calibration signs.
Innovation Solution
An online camera calibration method is introduced, utilizing calibration signs painted on the internal walls or doors of a garage facility, allowing for real-time calibration of camera angles without the need for manual intervention. Additionally, this method can be performed while the vehicle is in motion, using captured images of expected obstacles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If offline calibration using a mobile shelf with calibration signs is performed, then camera calibration accuracy is achieved, but calibration time is excessively long (up to 30 minutes) and manual intervention is required
Solution Approach 1:
The calibration system performs self-calibration automatically without manual intervention. The processor autonomously identifies calibration signs in images, calculates camera parameters, and updates calibration data, eliminating the need for operators to manually position mobile shelves and perform calculations
Solution Approach 2:
Calibration signs are pre-installed on fixed structures (walls, doors, poles) in the operating environment rather than using mobile shelves. This preliminary placement allows the calibration process to begin immediately when the vehicle enters the calibration area, eliminating time spent on manual setup and positioning
2Measurement precision
If offline calibration with mobile shelf is used, then camera calibration is completed, but the process requires manual intervention and operator involvement
Solution Approach 1:
The system achieves full automation through the processor that automatically captures images, identifies calibration sign features, calculates camera extrinsic and intrinsic parameters, and updates calibration databases without any manual intervention, transforming the calibration process from operator-dependent to fully autonomous
Solution Approach 2:
The manual mechanical process of positioning mobile shelves and manually recording calibration data is replaced by an automated vision-based system using image processing algorithms and computational geometry to calculate calibration parameters from captured images
3Measurement precision
If conventional calibration methods are used, then camera calibration is achieved, but the system cannot perform continuous calibration during vehicle operation
Solution Approach 1:
The calibration system transitions from a static offline process to a dynamic online process that can continuously update calibration parameters during vehicle operation. The system can capture images and recalculate calibration parameters in real-time as the vehicle moves through the calibration area, adapting to changes in camera positioning due to vibration and shock
Solution Approach 2:
The calibration process becomes continuous rather than periodic. The system can perform calibration at any time the vehicle passes through the calibration area, allowing for ongoing refinement of calibration parameters and eliminating the need for lengthy scheduled calibration sessions
Data Source
AI summary
In one embodiment, a system captures a first frame and a second frame for an environment of an autonomous driving vehicle (ADV) from at least a first and a second cameras mounted on the ADV. The system determines at least two points in the first frame having corresponding points in the second frame. The system determines distance and angle measurement information from the first camera to the at least two points and from the second camera to the corresponding points. The system determines actual positioning angles of the first and second cameras with respect to an orientation of the ADV based on the distance and angle measurement information and pixel information in the first and second frames. The actual positioning angles are used to compensate misalignments in positioning angles for the first and second cameras.