Camera Position Calculation Using Selective Correspondence Relationships
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional technologies face challenges in reducing the computational cost of calculating camera position and direction, as well as three-dimensional information from image features while maintaining accuracy, particularly in Structure from Motion (SfM) applications.
Innovation Solution
An information processing apparatus with a correspondence relationship acquisition unit, selection unit, and calculation unit that selects effective correspondence relationships based on specific indexes such as movement, number of images, similarity, distribution, and reliability to reduce the number of estimation parameters, thereby accelerating processing without compromising accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all correspondence relationships are used for calculation, then accuracy is maintained, but computational cost increases
Solution Approach 1:
The patent segments the set of all correspondence relationships into multiple groups based on spatial distribution and image overlap characteristics. By dividing the correspondence relationships into distinct groups (e.g., those with high image overlap vs. low image overlap), the system can selectively process only the most informative subsets, reducing computational load while preserving accuracy through strategic selection of representative correspondence relationships from each segment.
Solution Approach 2:
The patent applies local quality by evaluating and weighting different correspondence relationships based on their local characteristics such as image overlap ratio, spatial distribution, and reliability metrics. Instead of treating all correspondence relationships uniformly, the system identifies and prioritizes locally optimal correspondence relationships that contribute most to accuracy, thereby reducing the overall computational burden while maintaining measurement precision.
2Productivity
If the number of correspondence relationships is reduced, then processing speed increases, but accuracy may deteriorate
Solution Approach 1:
The patent implements preliminary action by pre-evaluating and ranking correspondence relationships based on multiple criteria including image overlap ratio, spatial distribution, and reliability indicators before the actual calculation process. This pre-screening step identifies and selects the most valuable correspondence relationships in advance, allowing the system to achieve high processing speed by processing only the pre-selected optimal subset while ensuring accuracy is maintained through the quality of the preliminary selection.
Solution Approach 2:
The patent utilizes parameter changes by dynamically adjusting selection thresholds and weighting parameters based on the specific characteristics of the input images and correspondence relationships. By changing parameters such as the minimum image overlap ratio threshold or the reliability weight factors, the system can optimize the balance between processing speed and accuracy for different scenarios, achieving high productivity without sacrificing measurement precision.
3Loss of information
If correspondence relationships with low image overlap are included, then completeness is improved, but computational complexity increases
Solution Approach 1:
The patent extracts and separately handles correspondence relationships with low image overlap by identifying them as a distinct category. Instead of processing all correspondence relationships uniformly, the system extracts only those with significant image overlap for primary processing, while using a simplified approach or selective inclusion for low-overlap cases. This extraction strategy maintains completeness by not discarding low-overlap relationships entirely, while reducing computational complexity by applying different processing levels.
Data Source
AI summary
An information processing apparatus according to an embodiment includes one or more hardware processors configured to function as a correspondence relationship acquisition unit, a selection unit, and a calculation unit. The correspondence relationship acquisition unit is configured to calculate a plurality of image features from a plurality of images captured by a camera and acquire a correspondence relationship between the image features. The selection unit is configured to select a plurality of correspondence relationships, based on effectiveness of the correspondence relationship and an influence on at least one of the images when the correspondence relationship with the effectiveness lower than an effectiveness threshold is eliminated. The calculation unit is configured to calculate at least one of a position and direction of the camera and three-dimensional information on the image features from the correspondence relationship selected from among the correspondence relationships.