Extrinsic Calibration of 2D and 3D LIDAR Sensors Using Point Cloud Entropy

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current methods for extrinsic calibration of 2D LIDAR sensors on vehicle platforms experiencing 3D motion are challenging due to reliance on known calibration targets and limited environmental coverage, and existing techniques are restricted to planar motion, making them unsuitable for complex environments.

Innovation Solution

An unsupervised algorithm using Renyi Quadratic Entropy as a point cloud quality metric to determine extrinsic calibration parameters between 2D and 3D LIDAR sensors, optimizing the quality of the 3D point cloud produced by the vehicle in an unknown environment, with a single tuning parameter and efficient computational methods for large datasets.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If known calibration targets and artificial calibration environments are used, then measurement precision is improved, but device complexity and ease of operation deteriorate due to requiring special targets and modified environments

Engineering Contradiction:

Improveextrinsic calibration precisionVSAvoidcalibration operation simplicity

Core Design Contradiction:

Measurement precisionVSEase of operation

Solution Approach 1:

The system performs self-calibration by using the environment itself as the calibration reference. The 3D LIDAR captures point clouds of the environment, and the 2D LIDAR scans are registered to this 3D reference without requiring any external calibration targets or artificial environments. The system serves its own calibration needs using readily available environmental data.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The 3D point cloud from the 3D LIDAR acts as an intermediary reference framework that mediates between the 2D LIDAR scans and the vehicle coordinate system. Instead of requiring direct measurement with complex targets, the 3D point cloud serves as a common reference that both sensors can be calibrated against through point cloud registration.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Device complexity

If 2D LIDAR is used for calibration, then device complexity is reduced, but measurement precision deteriorates due to limited coverage in 3D motion calibration

Engineering Contradiction:

Improvesensor system complexityVSAvoidcalibration accuracy in 3D motion

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The system merges the strengths of 3D LIDAR (full 360-degree coverage and 3D spatial information) with 2D LIDAR (simplicity and cost-effectiveness). The 3D LIDAR provides comprehensive environmental coverage to enable accurate calibration, while the 2D LIDAR maintains system simplicity. Their combined data is processed together through point cloud registration to achieve accurate extrinsic calibration.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The system transitions from 2D scan data to 3D point cloud representation to enable comprehensive calibration. By lifting the 2D LIDAR data into 3D space through registration with the 3D LIDAR point cloud, the system gains full 3D spatial coverage capability without requiring a 3D LIDAR for the calibration target itself.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

3Measurement precision

If supervised calibration procedures with multiple parameters are used, then measurement precision is improved, but loss of time increases due to lengthy computation

Engineering Contradiction:

Improvecalibration parameter accuracyVSAvoidcomputation time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system extracts only the essential extrinsic calibration parameters (position and orientation of the 2D LIDAR relative to the vehicle frame) from the full set of possible calibration parameters. By focusing specifically on these key parameters through point cloud registration, the system achieves accurate calibration without the computational burden of optimizing numerous intrinsic and extrinsic parameters simultaneously.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The system replaces traditional iterative optimization methods with point cloud registration techniques that leverage geometric constraints directly. Instead of using supervised procedures with arbitrary tuning parameters and lengthy iterative optimization, the system uses direct geometric matching between 2D scans and 3D point clouds, which converges faster and requires fewer tuning parameters.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP2761324B1Determining extrinsic calibration parameters for a sensor

Publication Date: 2020.02.19 THE CHANCELLOR MASTERS AND SCHOLARS OF THE UNIVERSITY OF OXFORD

Figure 1~2
Figure 3
Figure 4~5

AI summary

A method of determining extrinsic calibration parameters for at least one sensor (102, 104, 106) mounted on transportable apparatus (100). The method includes receiving (202) data representing pose history of the transportable apparatus and receiving (202) sensor data from at least one sensor mounted on transportable apparatus. The method generates (204) at least one point cloud data using the sensor data received from the at least one sensor, each point in a said point cloud having a point covariance derived from the pose history data. The method then maximises (206) a value of a quality function for the at least one point cloud, and uses (208) the maximised quality function to determine extrinsic calibration parameters for the at least one sensor.