Nvidia patents method to auto-calibrate sensors using object shapes
Nvidia describes a method for calibrating sensors by comparing the shapes of objects detected in images to expected shapes. The system uses an optimization process to adjust calibration parameters until the detected shapes match the expected ones closely enough.
In the filing’s own words
In various examples, systems and methods are provided for parameter generation for sensor calibration. A dataset of samples bounding shape(s) may be established from detection bounding shape(s) corresponding to object(s) detected in image(s) represented by sensor data captured by a sensor. A parameter space may be searched for estimated calibration parameters for the sensor using an optimization process. The optimization process may include iteratively generating sets of candidate calibration parameters for the sensor based at least on a difference between sets of candidate bounding shapes generated from the sets of candidate calibration parameters and the sampled bounding shape(s). The optimization process may include outputting the estimated calibration parameters for the sensor in response to a difference between a set of candidate bounding shapes corresponding to the estimated calibration parameters for the sensor and the sampled bounding shape(s) satisfying a first threshold and/or level of convergence.
What exactly they claim (claim 1)
1. One or more processors comprising processing circuitry to: generate a set of estimated calibration parameters for a sensor by: iteratively generating one or more sets of candidate calibration parameters for the sensor; generating one or more sets of candidate bounding shapes, wherein each set of the one or more sets of candidate bounding shapes corresponds to a respective set of the one or more sets of candidate calibration parameters and one or more sampled bounding shapes corresponding to one or more objects detected in one or more images, the one or more images representing sensor data captured by the sensor; determining a difference between the one or more sets of candidate bounding shapes and the one or more sampled bounding shapes; and determining the set of estimated calibration parameters for the sensor in response to a difference between a set of candidate bounding shapes corresponding to the set of estimated calibration parameters and the one or more sampled bounding shapes satisfying one or more of a first threshold or a level of convergence; and cause the sensor to be calibrated using the set of estimated calibration parameters.
Facts
- Company
- Nvidia (NVIDIA Corporation)
- Official title
- PARAMETER GENERATION FOR SENSOR CALIBRATION
- Publication number
- US20260268524A1
- Application number
- 19072646
- Filed
- Mar 6, 2025
- Published
- Sep 10, 2026
- Inventors
- 3
- Claims
- 20
Classified under
G06T 7/80 Physics & computingG06T 7/11 Physics & computingG06T 7/70 Physics & computingG06V 10/25 Physics & computingG06V 2201/07 Physics & computingMore from Nvidia in AI & software
- Nvidia patents AI coach that analyzes your gameplay to suggest strategies Oct 13, 2026
- Nvidia patents AI system that writes and voices character conversations Sep 10, 2026
- Nvidia patents system that uses AI to find bugs in game videos Sep 17, 2026
- Nvidia patents AI method to make digital characters move realistically Sep 8, 2026
- Nvidia patents AI system that builds image generators from text prompts Sep 15, 2026
Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.