Auto-Calibrating Single Camera Using Virtual Object Key Points

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current video monitoring systems with single cameras face inefficiencies in calibrating 2D image data to 3D mapping, requiring labor-intensive and costly manual processes that are prone to errors, and necessitate repeated calibration due to camera pose changes or power loss.

Innovation Solution

The system automatically calibrates a single camera using data-driven techniques, analyzing images to identify objects, generating virtual objects, and reprojecting key points to fit the 2D space, allowing for self-monitoring and self-recovery without human intervention, using a flat earth camera model to determine reprojection parameters for accurate 3D mapping.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual calibration process is used with known pattern, then camera calibration accuracy is achieved, but labor cost and time consumption increase significantly

Engineering Contradiction:
Improvecamera calibration accuracyVSAvoidcalibration time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs automatic calibration without human intervention by detecting objects in the scene and using their inherent geometric features. The calibration process serves itself by utilizing naturally occurring objects (vehicles, pedestrians, infrastructure) rather than requiring external calibration patterns, thereby eliminating manual placement and measurement time while maintaining accuracy through algorithmic object recognition and feature extraction.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system creates a virtual 3D model of detected objects and uses this digital copy to establish correspondence between 2D image coordinates and 3D world coordinates. By replicating object geometry in virtual space and performing transformations on the copy, the system avoids repeated manual measurements of physical objects while preserving calibration accuracy through mathematical modeling.

Inventive Principle:
Principle #26Copying

2Measurement precision

If manual calibration process is used with known pattern, then camera calibration accuracy is achieved, but manpower cost increases

Engineering Contradiction:
Improvecamera calibration accuracyVSAvoidautomation level
Core Design Contradiction:
Measurement precisionVSExtent of automation

Solution Approach 1:

The calibration system operates autonomously by automatically detecting objects in the camera field of view, extracting their geometric features, and computing calibration parameters without human assistance. The system self-monitors calibration quality and performs self-recovery when pose changes are detected, completely eliminating the need for manual pattern placement and measurement that characterized traditional calibration methods.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system replaces the mechanical manual process of placing and measuring physical calibration patterns with an automated computer vision-based approach. Instead of human operators physically positioning known patterns and manually measuring their coordinates, the system uses image processing algorithms to automatically detect objects and compute their 3D positions from 2D images, substituting mechanical human labor with computational automation.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Measurement precision

If traditional calibration process is used, then accurate 2D to 3D mapping is achieved, but repeated calibration is required due to camera pose changes

Engineering Contradiction:
Improve2D to 3D mapping accuracyVSAvoidcalibration efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The system continuously monitors the camera pose and detects changes in the environment. When a pose change is detected, the system automatically triggers a re-calibration process using newly detected objects. This feedback mechanism ensures that calibration accuracy is maintained without requiring manual intervention, as the system self-adjusts by detecting objects and recomputing calibration parameters when needed.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The system performs preliminary object detection and feature extraction in advance to prepare calibration data. By continuously analyzing the scene for detectable objects and pre-computing their geometric features, the system is ready to perform rapid calibration when pose changes occur, eliminating the need for time-consuming manual pattern placement and enabling quick automated recalibration.

Inventive Principle:
Principle #10Preliminary action

4Measurement precision

If manual pattern placement is used for calibration, then calibration accuracy is achieved, but error-prone human involvement is required

Engineering Contradiction:
Improvecalibration accuracyVSAvoidcalibration reliability
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The calibration system eliminates human involvement entirely by automatically detecting objects and computing calibration parameters. The system serves itself by using algorithmic object recognition and automated coordinate transformation, removing the human error sources associated with manual pattern placement, measurement, and data entry while maintaining calibration accuracy through robust computational methods.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system creates virtual 3D copies of detected objects and uses these digital representations for calibration computations. By working with replicated geometric models rather than physical patterns, the system eliminates errors from manual pattern handling and measurement, achieving reliable and repeatable calibration results through consistent mathematical transformations of virtual object coordinates.

Inventive Principle:
Principle #26Copying

Data Source

PatentEP4168989B1Auto calibrating a single camera from detectable objects
Publication Date: 2024.05.22 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP4168989B1 patent drawingFigure 1
  • EP4168989B1 patent drawingFigure 2
  • EP4168989B1 patent drawingFigure 3

AI summary

Techniques for improved camera calibration are disclosed. An image is analyzed to identify a first set of key points for an object. A virtual object is generated. The virtual object has a second set of key points. A reprojected version of the second set is fitted to the first set in 2D space until a fitting threshold is satisfied. To do so, a 3D alignment of the second set is generated in an attempt to fit (e.g., in 2D space) the second set to the first set. Another operation includes reprojecting the second set into 2D space. In response to comparing the reprojected second set to the first set, another operation includes determining whether a fitting error between those sets satisfies the fitting threshold. A specific 3D alignment of the second set is selected. The camera is calibrated based on resulting reprojection parameters.