Feature Point Matching Control for Real-Time Visual SLAM
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods for feature point matching in SfM and vSLAM require significant computational complexity, leading to long processing times, and previous methods to reduce complexity either fail to complete in real-time or do not sufficiently decrease processing amounts.
Innovation Solution
An image processing apparatus that selects reference and search subject images based on the movement direction of a movable apparatus, limiting the search range for feature point association, and uses a pre-determined search range map to reduce computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the search range is limited based on self-position estimation information or vanishing point position, then computational complexity of feature point matching is decreased, but processing time cannot be reduced sufficiently or processing amount cannot be sufficiently decreased
Solution Approach 1:
The patent pre-determines search range maps based on possible movement directions (up, down, left, right) before actual feature point matching occurs. These search range maps are stored in advance, eliminating the need for complex real-time calculations of search ranges during matching operations. This preliminary preparation resolves the contradiction by reducing computational complexity through pre-computation while maintaining high processing speed through efficient lookup and application of pre-determined ranges.
2Reliability
If corresponding points are searched for from all input images, then comprehensive feature point matching is achieved, but vast computational complexity becomes necessary and processing time becomes long
Solution Approach 1:
The patent applies local quality by making the search range specific to each feature point based on its location and the determined movement direction. Instead of using a uniform search range across all images, the search range is locally optimized for each feature point by referencing pre-determined search range maps. This ensures comprehensive matching within relevant regions while avoiding unnecessary computations in irrelevant areas, thus maintaining matching accuracy while reducing computational complexity.
3Reliability
If the search range is set from the vanishing point to the edge of the image, then the search covers all possible features, but the processing amount cannot be sufficiently decreased
Solution Approach 1:
The patent introduces dynamics by making the search range adaptive to the determined movement direction. Instead of using a static search range from vanishing point to edge, the system dynamically selects and applies appropriate search range maps based on the detected movement direction (up, down, left, or right). This dynamic adjustment ensures search completeness in the relevant direction while significantly reducing the processing amount by excluding irrelevant regions, thus resolving the contradiction between search completeness and processing speed.
Data Source
AI summary
An image processing apparatus includes a memory storing instructions; and a processor executing the stored instructions causing the image processing apparatus to: select, from plural time series images acquired from an image capturing apparatus disposed on a movable apparatus capturing surrounding images, a first image and a second image captured when the movable apparatus was at a first position and a second position, respectively, and control a movement direction of an image capturing position to the first and second positions; select a reference image and a search subject image from the first image and the second image, respectively, and perform feature point association; and switch the reference image based on information showing a movement direction of the movable apparatus. During the feature point association, each feature point in the search subject image is searched for within a search range limited based on the movement direction of the movable apparatus.