Self-Location Estimation Using Cross-Condition Feature Association
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional self-location estimation technologies face challenges in improving accuracy when various environmental factors such as seasonal changes, weather, and lighting conditions affect image appearance, leading to a reduction in usable features and degraded estimation accuracy.
Innovation Solution
An estimation device that includes a processing unit with query and reference image acquisition units, feature calculation units, and a self-position calculation unit, which acquires and processes images and imaging conditions to calculate features and similarities, allowing for accurate self-location estimation by associating reference images with query images based on shared features despite varying conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional self-location estimation technology is used, then the system can estimate photographing position using image databases, but the accuracy degrades when environmental factors such as seasonal changes, weather, and lighting conditions vary
Solution Approach 1:
The patent introduces imaging condition information (weather, season, time, lighting) as an intermediary element that mediates between the query image and reference image comparison. This intermediary allows the system to account for environmental variations by selecting or weighting reference images with similar imaging conditions, thereby maintaining high estimation accuracy across diverse environmental conditions without requiring the system to adapt to each condition separately
Solution Approach 2:
The patent changes the parameter set used for image comparison by incorporating imaging condition parameters (weather, season, time of day, lighting conditions) alongside traditional visual features. This multi-parameter approach allows the system to maintain high measurement precision by matching both the visual content and the environmental context of images, effectively resolving the contradiction between accuracy and environmental adaptability
2Measurement precision
If pixel-wise correspondence and geometric calculation are used to improve estimation accuracy, then the calculation complexity and processing time increase significantly
Solution Approach 1:
The patent segments the self-location estimation process into multiple independent stages: first extracting visual features from images, then extracting imaging condition features, followed by similarity calculation using both feature types, and finally position estimation. This segmentation allows each stage to be optimized independently and enables parallel processing, reducing overall calculation complexity while maintaining high precision through the combined use of multiple feature types
Solution Approach 2:
The patent employs partial action by selectively using different feature types and matching strategies based on the specific application scenario and available computational resources. The system can adjust the depth of feature extraction and similarity calculation, performing only the necessary computations to achieve the required accuracy level, thereby avoiding excessive calculation complexity while maintaining sufficient estimation precision
3Measurement precision
If multiple feature types are extracted and combined for image association, then the processing time and computational load increase
Solution Approach 1:
The patent applies preliminary action by pre-extracting and storing imaging condition information (weather, season, time, lighting conditions) along with reference images in the database during the database construction phase. This preliminary preparation allows the system to quickly retrieve and compare both visual and environmental features during query processing, significantly reducing real-time processing time while maintaining high association accuracy through multi-feature comparison
Data Source
AI summary
An estimation device according to an embodiment includesone or more hardware processors configured to function as a query-image-acquisition unit, a query-imaging-condition-acquisition unit, a reference-image-acquisition unit, a reference-imaging-condition-acquisition unit, a feature-amount calculation unit, and a self-position calculation unit. The query-imaging-condition-acquisition unit acquires an imaging condition of a query image. The reference-imaging-condition-acquisition unit acquires an imaging condition of a reference image. The feature-amount calculation unit calculates a query image's feature based on the reference imaging condition, and calculates a reference image's feature based on the query imaging condition. The self-position calculation unit associates the reference image with the query image based on the query feature and the reference feature, and uses a position/pose of a second imaging device capturing the reference image associated with the query image to calculate a self location indicating at least one of a position/pose of a first imaging device.