3D Map Target Tracking via Sub-Space Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional SLAM technology for map target tracking in 3D spaces requires significant resources and time due to the need for comparing all key frames and landmarks, limiting efficient location recognition in continuous 3D spaces.
Innovation Solution
Divide the 3D space into sub-spaces, acquiring sub-images, creating sub-maps with main key frames, and tracking posture information using a 3D main map, reducing data processing by comparing only key frames and feature points within sub-spaces.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all key frames and landmarks in the pre-created SLAM map are compared for location tracking, then location recognition accuracy is improved, but processing time and resource consumption increase significantly
Solution Approach 1:
The patent divides the continuous 3D space into multiple discrete sub-spaces, each represented by a key frame. Instead of comparing the current image with all key frames in the entire map, the system only compares with key frames within the relevant sub-space, significantly reducing the number of comparisons while maintaining location recognition accuracy.
Solution Approach 2:
The patent extracts and utilizes only the necessary key frames and landmarks from the pre-created SLAM map that are relevant to the current location, rather than processing all key frames and landmarks. This extraction of essential elements reduces data processing requirements while preserving tracking accuracy.
2Measurement precision
If all key frames and landmarks in the pre-created SLAM map are compared for location tracking, then location recognition accuracy is improved, but computational resources increase significantly
Solution Approach 1:
The patent segments the SLAM map into multiple sub-spaces, allowing the system to process only the subset of key frames and landmarks relevant to the current sub-space. This segmentation reduces the computational burden by limiting the scope of comparison operations to only necessary data elements.
Solution Approach 2:
The patent applies partial action by performing comparison operations only on the necessary key frames and landmarks within the relevant sub-space, rather than exhaustively comparing all elements in the map. This partial processing approach reduces computational resource consumption while achieving sufficient tracking accuracy.
3Productivity
If the pre-created SLAM map is divided at predetermined key frame intervals, then processing efficiency is improved, but it becomes difficult to reach location recognition in meaningful spatial units
Solution Approach 1:
The patent applies local quality by creating sub-spaces based on meaningful spatial criteria rather than uniform key frame intervals. Each sub-space is defined to represent a coherent spatial unit, ensuring that location recognition occurs at meaningful spatial scales while still improving processing efficiency through localized comparisons.
Data Source
AI summary
A method of tracking a map target according to one embodiment of the present disclosure, which tracks the map target through a map target tracking application executed by at least one processor of a terminal, includes: acquiring a basic image obtained by photographing a 3D space; acquiring a plurality of sub-images obtained by dividing the acquired basic image for respective sub-spaces in the 3D space; creating a plurality of sub-maps based on the plurality of acquired sub-images; determining at least one main key frame for each of the plurality of created sub-maps; creating a 3D main map by combining the plurality of sub-maps for which the at least one main key frame is determined; and tracking current posture information in the 3D space based on the created 3D main map.