Vehicle Sensor Fusion for Automatic Camera Parameter Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face errors in distance estimation due to changes in camera external parameters caused by shocks or shaking, and conventional correction methods require manual calibration, which is inconvenient.
Innovation Solution
A vehicle system that includes image sensors, radar/lidar sensors, and a controller to process image and sensing information, determine if external parameter correction is necessary based on distance errors, and update the parameters automatically to maintain accurate distance estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual calibration is performed to correct camera external parameters, then measurement precision is improved, but ease of operation deteriorates
Solution Approach 1:
The system performs automatic external parameter correction using onboard sensors (radar, lidar, gyroscope) and image processing algorithms. The controller automatically detects camera position shifts by comparing sensor data with image information and corrects external parameters without requiring manual intervention, making the system self-correcting while maintaining high distance estimation accuracy
Solution Approach 2:
The system continuously monitors distance values from multiple sensors and compares them with image-based distance estimates. When discrepancies exceed a threshold, the system triggers automatic correction of camera external parameters based on gyroscope data and sensor fusion, creating a closed-loop feedback mechanism that maintains measurement precision without manual calibration
2Ease of operation
If automatic correction is implemented to improve ease of operation, then ease of operation is improved, but device complexity increases
Solution Approach 1:
The controller performs multiple functions: it processes image information from the camera, fuses data from radar and lidar sensors, monitors distance estimates, detects parameter drift, and executes correction algorithms. By making the controller multi-functional, the system avoids adding separate dedicated hardware for each function, thereby managing complexity while providing automatic correction capabilities
Solution Approach 2:
The system combines data from multiple sensors (camera, radar, lidar, gyroscope) and integrates their processing functions into a unified correction algorithm. The external parameter correction uses combined information from sensor fusion and image processing, merging multiple subsystems into a coordinated automatic correction mechanism that reduces overall system complexity
3Measurement precision
If frequent calibration is performed to maintain measurement precision, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The system dynamically adjusts the calibration/correction frequency based on detected camera position shifts and environmental conditions. Instead of fixed frequent calibration, the system performs corrections only when drift is detected beyond thresholds, adapting the correction schedule to actual system behavior and reducing unnecessary time loss while maintaining precision when needed
Data Source
AI summary
A vehicle includes a first sensor provided to have a field of view facing the surroundings of the vehicle to generate image information, a second sensor including at least one of a radar sensor or a lidar sensor to generate sensing information about the surroundings of the vehicle, and a controller. The controller is configured to identify an object around the vehicle based on processing of the image information and the sensing information, identify distances between the identified surrounding object and the vehicle based on each of the image information and the sensing information, and determine whether correction of an external parameter of the first sensor is necessary based on each of the identified distances and the reference error distribution information.