SLAM Map Updating with Repetitive Pattern Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing SLAM systems face challenges in accurately constructing maps of wide environments due to limitations in camera image quality, leading to inefficient and low-accuracy map construction, particularly when capturing distant objects.
Innovation Solution
An electronic device employs image recognition to identify repetitive patterns in key frames, generates map points corresponding to these patterns, and updates the map using inferred map points to enhance accuracy and completeness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If camera image quality is used for map construction in wide environments, then the system can operate with simple equipment, but the accuracy of distant object features deteriorates due to inherent camera limitations
Solution Approach 1:
The patent creates virtual copies of map points by inferring their positions based on repetitive pattern recognition. When a repetitive pattern is detected, the system generates multiple inferred map points corresponding to each instance of the pattern, effectively copying the positional information from one detected instance to multiple instances. This allows accurate representation of distant objects without requiring high-resolution camera images.
Solution Approach 2:
The patent introduces an intermediary processing step between image capture and map construction. Instead of directly using camera images for map points, the system first performs image recognition to identify repetitive patterns, then uses these patterns as intermediaries to infer map point positions. This intermediary process compensates for camera limitations by leveraging pattern repetition information.
2Ease of manufacture
If traditional SLAM methods are used for map construction, then the system implementation is straightforward, but the map construction efficiency deteriorates due to insufficient feature detection in wide environments
Solution Approach 1:
The patent performs preliminary image recognition to identify repetitive patterns before proceeding with map point generation. By detecting repetitive patterns in advance, the system can pre-calculate inferred map points for multiple instances of the pattern, significantly improving map construction efficiency. This preliminary action allows the system to handle wide environments more effectively by preparing feature information before full map construction.
3Ease of operation
If camera imaging is used for distant objects, then the system maintains simple operation, but the map accuracy deteriorates due to image quality limitations
Solution Approach 1:
The system creates accurate representations of distant objects by copying positional information from detected repetitive patterns. When a pattern is identified, the system generates multiple inferred map points that accurately represent the positions of distant instances, overcoming camera resolution limitations while maintaining simple camera-based operation.
Solution Approach 2:
The patent changes the approach from direct image measurement to pattern-based inference. Instead of relying on camera image parameters (resolution, distance), the system uses pattern recognition parameters (repetition count, pattern consistency) to determine map point positions, thereby improving accuracy without changing the camera or operation method.
Data Source
AI summary
The present disclosure provides a map updating method and an electronic device. The map updating method is applicable to the electronic device configured to establish a map of an environment, and includes: performing an image recognition on a key frame, to determine if the key frame comprises a repetitive pattern; obtaining a plurality of map points corresponding to the repetitive pattern; generating a plurality of inferred map points according to the plurality of map points; and updating the map according to the plurality of inferred map points.