Camera Self-Calibration Using Scene-Based Feature Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Distortion occurs in captured images due to impacts or camera performance deterioration, causing subjects to be stored in shapes different from their actual appearance.
Innovation Solution
An electronic apparatus that extracts feature points based on image type and corrects camera parameters using a processor to identify scene types, assign reliability scores, and update parameters based on integrated calibration values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If camera parameters are corrected using traditional calibration methods, then manufacturing precision is improved, but device complexity increases due to requiring complex calibration procedures and multiple captured images
Solution Approach 1:
The system performs self-calibration by automatically identifying scene types and extracting feature points from captured images without requiring manual intervention or complex external calibration equipment. The processor autonomously updates camera parameters based on analyzed image data, enabling the device to self-correct distortion issues.
Solution Approach 2:
The system changes the approach from traditional complex calibration to a simplified parameter update method. By identifying scene types and selecting appropriate feature point extraction methods, the system dynamically adjusts calibration parameters (focal length, principal point, skew coefficient, distortion coefficient) based on the specific characteristics of captured images, achieving accurate correction with simpler procedures.
2Measurement precision
If multiple captured images are used for calibration, then measurement precision is improved, but loss of time increases due to processing multiple images
Solution Approach 1:
The calibration process is segmented by dividing images into different scene types (outdoor, indoor, regular pattern). Each scene type has its own optimized feature point extraction method, allowing the system to process images more efficiently by applying the most suitable method for each category rather than using a single complex method for all images.
Solution Approach 2:
The system extracts only the necessary feature points required for calibration based on the identified scene type, rather than processing all possible image features. This partial action approach focuses computational resources on the most relevant features for each scene type, reducing overall processing time while maintaining calibration accuracy.
3Manufacturing precision
If feature points are extracted based on scene type, then manufacturing precision is improved, but device complexity increases due to requiring scene type identification
Solution Approach 1:
The system dynamically adapts its feature point extraction method based on the identified scene type. Rather than using a static, one-size-fits-all approach, the processor selects and applies different extraction algorithms appropriate for outdoor scenes, indoor scenes, or regular pattern scenes, optimizing accuracy for each context while maintaining a manageable system structure.
Data Source
AI summary
An electronic apparatus is provided. The electronic apparatus includes a camera, a memory in which a plurality of captured images obtained through the camera and a parameter value of the camera are stored, and a processor electrically connected to the camera and the memory. The processor is configured to identify a scene type corresponding to each captured image of the plurality of captured images; extract a feature point of each captured image of the plurality of captured images based on a feature point extraction method corresponding to the identified scene type of each captured image; obtain a calibration parameter value corresponding to a feature type of each extracted feature point; obtain an integrated calibration parameter value based on one or more obtained calibration parameter values; and update the parameter value stored in the memory based on the integrated calibration parameter value.