LIDAR-Camera Calibration Using Board Plane Transformation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional maps for autonomous vehicles lack precision and accuracy, leading to challenges in safe navigation due to outdated data and high costs associated with creating and maintaining high-definition maps, which can result in inaccurate location determination and delayed updates of road conditions.
Innovation Solution
The development of a system that generates and maintains high-definition (HD) maps using data from autonomous vehicles' sensors, allowing for real-time updates and precise location tracking, while also enabling sensor calibration using boards with checkerboard or diamond patterns to improve data correlation and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional maps are used for autonomous vehicle navigation, then the system is simpler and cheaper to implement, but the location determination accuracy and safety threshold are insufficient
Solution Approach 1:
The patent combines multiple data sources including LIDAR sensor data, camera images, and map data into a unified coordinate system through sensor calibration. This integration allows the system to achieve high-precision location determination (within 30 cm safety threshold) by merging the strengths of different sensing modalities rather than relying on a single map system.
Solution Approach 2:
The patent introduces sensor calibration as an intermediary process that transforms and aligns data from different sensors (LIDAR, camera) with map coordinates. This calibration mechanism serves as a mediator that reconciles the differences between sensor measurements and map representations, enabling accurate location determination without requiring the vehicle to carry complex dedicated positioning hardware.
2Measurement precision
If survey teams create comprehensive HD maps using specialized survey cars, then map accuracy is improved, but the cost and time required to create and update maps increases significantly
Solution Approach 1:
The patent enables autonomous vehicles to perform their own sensor calibration using a calibration board with checkerboard or diamond patterns that they encounter during normal operation. This self-calibration capability eliminates the need for specialized survey teams and expensive survey cars, allowing vehicles to autonomously maintain accurate sensor-to-map transformations while driving.
Solution Approach 2:
The system implements a feedback mechanism where calibration data collected from the vehicle's own sensor measurements against known calibration board patterns is used to continuously refine and update the sensor-to-map transformation parameters. This feedback loop allows the system to maintain high map accuracy without requiring external survey interventions, thereby increasing map update frequency to match road change rates.
3Reliability
If conventional map updating methods are used, then the map data remains stable and consistent, but the data becomes outdated and fails to reflect current road conditions
Solution Approach 1:
The patent establishes continuous sensor calibration and map updating as an ongoing process that occurs whenever the autonomous vehicle operates and encounters calibration boards. This continuous calibration approach ensures map data remains fresh and reflects current road conditions without requiring periodic interruptions for specialized survey operations, thereby maintaining both reliability and data freshness simultaneously.
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 system provides accurate and up-to-date HD maps for safe navigation, enabling autonomous vehicles to operate within a target safety threshold with improved sensor calibration, reducing latency and storage requirements, and allowing for efficient map management.
Implementation Method 1
receiving a LIDAR scan of a scene from a LIDAR of a vehicle
Implementation Method 2
detecting the board in a camera image from a camera of the vehicle
Data Source
AI summary
According to an aspect of an embodiment, operations may comprise receiving a LIDAR scan of a scene from a LIDAR of a vehicle with the scene comprising a board, detecting the board in the LIDAR scan, fitting a plane through LIDAR coordinates corresponding to the detected board, projecting the plane from the LIDAR coordinates to a first set of camera coordinates, detecting the board in a camera image from a camera of the vehicle at a second set of camera coordinates, and calibrating the LIDAR of the vehicle and the camera of the vehicle by determining a transform between the first set of camera coordinates and the second set of camera coordinates.