Camera Device Localization Using Sparse Visual Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for localizing camera-equipped devices using computer vision are costly due to the processing-intensive search of three-dimensional maps of visual features, which can be mitigated by using a sparse map and delegating tasks between wireless access points.
Innovation Solution
A system that enables communication between a wireless access point and a camera-equipped device to determine the device's point of view within a local three-dimensional map of visual features, reducing the search space using non-visual parameters and delegating tasks to other access points if localization is unsuccessful.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a dense three-dimensional map of visual features is used for device localization, then localization accuracy is improved, but processing cost and memory requirements increase
Solution Approach 1:
The patent divides the large-scale three-dimensional map into multiple local three-dimensional maps, each covering a specific physical area associated with a wireless access point. This segmentation reduces the search space from a global map to localized regions, significantly decreasing processing requirements while maintaining localization accuracy within each local area.
Solution Approach 2:
The patent creates local three-dimensional maps with high visual feature density only in specific physical areas where wireless access points are deployed. Each local map is optimized for its specific region, providing high localization accuracy locally while avoiding the computational burden of maintaining a dense global map.
2Measurement precision
If a dense three-dimensional map of visual features is used for device localization, then localization accuracy is improved, but memory requirements increase
Solution Approach 1:
The patent segments the global map into multiple local three-dimensional maps stored at different wireless access points. Each access point stores only the local map for its coverage area, distributing memory requirements across the network infrastructure rather than requiring one device to hold the entire global map.
Solution Approach 2:
The patent extracts only the necessary local portion of the three-dimensional map relevant to the device's current physical area. Instead of loading or processing the entire global map, the system retrieves and processes only the local map data needed for current localization, significantly reducing memory usage.
3Productivity
If task delegation between wireless access points is implemented, then processing efficiency is improved, but system complexity increases
Solution Approach 1:
The patent combines the localization functionality across multiple wireless access points into a coordinated system. Each access point performs localization using its local map, and unsuccessful localizations are automatically delegated to neighboring access points, merging their computational resources to achieve high processing efficiency without requiring complex centralized control.
Solution Approach 2:
The patent implements a self-service mechanism where each wireless access point autonomously attempts localization using its own local map first. Only when localization fails does the system delegate to neighboring access points, allowing each node to serve itself before seeking external help, thereby reducing overall system complexity while maintaining high productivity.
Data Source
Figure 1~5
Figure 6~7
Figure 8~9
AI summary
An apparatus comprising: means for enabling communication between a wireless access point to a network and a camera-equipped wireless client device; and means for enabling determination of a particular point of view that corresponds to a visual image captured by the camera-equipped wireless client device, to localize the camera-equipped wireless client device, within a physical area covered, for wireless access, by the wireless access point, wherein the determination uses data corresponding to visual features associated with the physical area covered, for wireless access, by the wireless access point.