Lidar Camera Extrinsic Calibration Using Chessboard Target
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for calibrating lidar and camera sensors require extensive, expensive, and time-consuming processes in controlled environments, limiting their ability to be used in autonomous navigation and requiring frequent recalibration due to sensor misalignment.
Innovation Solution
A method and system for establishing an extrinsic relationship between a lidar sensor and a camera using a target chessboard, which allows for rapid calibration in the field by determining translation and rotation parameters through optimization algorithms, enabling the fusion of sensor data for autonomous systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional Metrologic methods are used for extrinsic calibration, then measurement precision is improved, but loss of time and productivity deteriorate due to the expensive and time-consuming process requiring controlled environments and meticulous staging
Solution Approach 1:
The system uses the autonomous vehicle's own sensors (lidar and camera) to perform self-calibration without requiring external surveying equipment or controlled environments. The calibration target is imaged by the vehicle's native sensors, enabling the system to calibrate itself in real-world conditions rather than requiring specialized measurement infrastructure
Solution Approach 2:
A calibration target with machine-readable pattern (such as a chessboard) serves as an intermediary object that can be easily imaged by both lidar and camera sensors. This target provides common reference features that enable the computation of extrinsic calibration parameters between different sensor modalities without requiring complex Metrologic equipment
2Measurement precision
If traditional extrinsic calibration methods are used, then measurement precision is improved, but device complexity worsens due to requirements for disassembling platform components and accessing sensors in confined spaces
Solution Approach 1:
The calibration procedure requires no disassembly or special access to sensor components. The autonomous vehicle simply needs to capture images of the calibration target with its existing sensors, making the process as simple as taking a photograph rather than requiring technical intervention with the sensor mounting structure
Solution Approach 2:
The calibration process extracts only the essential information needed for calibration from the imaged target (such as corner positions of a chessboard pattern), separating the calibration function from complex physical measurement procedures. This extraction approach simplifies the process to basic image capture and computational geometry
3Ease of operation
If sensors are used in isolation without calibration, then ease of operation is improved, but reliability deteriorates because meaningful fusion of information from multiple sensors cannot be achieved
Solution Approach 1:
The calibration process merges the coordinate systems of multiple sensors (lidar and camera) into a common reference frame through computed extrinsic parameters. This merging enables seamless fusion of sensor data while maintaining the operational simplicity of using multiple sensors together, as the calibration establishes the mathematical relationship needed for accurate integration
Data Source
AI summary
In one or more embodiments, a method for calibration between a lidar sensor and a camera comprises determining translation parameters of extrinsic calibration parameters by using a location of the camera with respect to the lidar sensor. The method further comprises orienting a target chessboard such that it is aligned with axes of a lidar coordinate system. Also, the method comprises optimizing a best fit transformation between a camera coordinate system and a target chessboard coordinate system. In addition, the method comprises determining a rotation matrix using the best fit transformation between the camera coordinate system and the target chessboard coordinate system. Additionally, the method comprises extracting Euler angles of the extrinsic calibration parameters from the rotation matrix. Also, the method comprises collecting, by the lidar sensor and the camera, calibration scenes using the target chessboard. Further, the method comprises optimizing the extrinsic calibration parameters by using the calibration scenes.