Camera-LiDAR Extrinsic Validation for Sensor Miscalibration Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Vehicle sensors, such as cameras and lidars, can become miscalibrated due to environmental factors or internal errors, leading to suboptimal or unsuitable sensor data that impacts vehicle navigation, obstacle detection, and other critical functions.
Innovation Solution
A computing device implements a method to determine the probability of sensor miscalibration by generating perturbed extrinsics based on calibrated extrinsics, calculating costs for these perturbed extrinsics, and determining a local maxima score to assess the likelihood of miscalibration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sensor calibration is performed using traditional methods, then calibration can be completed, but the reliability of calibration validation is insufficient and miscalibration may go undetected
Solution Approach 1:
The system performs preliminary validation by generating perturbed extrinsics before final calibration acceptance. This preliminary action involves creating multiple slightly varied calibration parameters and testing them in advance to determine if the original calibration is robust, thereby preventing undetected miscalibration from propagating into the navigation system.
Solution Approach 2:
The system implements feedback by calculating a validation metric based on comparing costs of perturbed extrinsics against the original calibration. This feedback mechanism provides quantitative information about calibration quality, allowing the system to detect miscalibration and trigger recalibration when the validation metric falls below a threshold.
2Measurement precision
If perturbed extrinsics are generated with fine step sizes, then miscalibration detection precision is improved, but computational complexity increases
Solution Approach 1:
The system applies partial action by generating a limited set of perturbed extrinsics with small step sizes (e.g., 0.1 to 1 degree) rather than exhaustively searching all possible calibration parameters. This partial exploration is sufficient to detect miscalibration while avoiding excessive computational burden that would result from comprehensive parameter space search.
Data Source
AI summary
Techniques for determining a probability that a first sensor is miscalibrated with respect a second sensor are discussed herein. For example, a computing device may receive calibrated extrinsics of a camera to a lidar, determine a plurality of sets of perturbed extrinsics based on the calibrated extrinsics, determine respective costs for perturbed extrinsics of the plurality of sets of perturbed extrinsics based on image data captured by the camera, the plurality of sets of perturbed extrinsics, and lidar data captured by the lidar, and determine a local maxima score for the calibrated extrinsics based at least in part on the respective costs for the perturbed extrinsics of the plurality of sets of perturbed extrinsics and a cost of the calibrated extrinsics. The computing device may then determine a probability that the camera is miscalibrated based on a Bayes probability and the local maxima score.