Place Recognition Algorithm Using Two-Stage ORB Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional SLAM techniques face challenges in real-time camera pose determination due to the use of sophisticated feature descriptors and brute-force feature matching, which are inefficient and fail to account for rotation and viewpoint changes effectively.
Innovation Solution
A place recognition algorithm that extracts a smaller subset of candidate key frames and performs pair-wise matching using a two-stage process, relying solely on image content for key frame addition, enabling real-time camera pose determination by employing ORB binary features and a hierarchical bag-of-words model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional sophisticated feature descriptors and brute-force feature matching are used, then measurement precision is improved, but productivity deteriorates
Solution Approach 1:
The patent segments the feature matching process into two stages: first extracting distinctive keypoints and descriptors, then performing matching only on extracted features rather than brute-force comparison of all image pixels. This segmentation enables real-time processing while maintaining accuracy
Solution Approach 2:
The patent replaces traditional mechanical brute-force feature matching with a more efficient algorithmic approach using ORB (Oriented FAST and Rotated BRIEF) descriptors and hierarchical matching, substituting computational mechanics for exhaustive search methods
2Measurement precision
If traditional feature matching is used, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The patent performs preliminary actions by pre-extracting and storing ORB descriptors for keyframes before actual matching occurs. This preliminary preparation enables rapid real-time matching without repeated heavy computation, reducing processing time while maintaining precision
Solution Approach 2:
The patent changes parameters by using binary ORB descriptors instead of traditional float-based descriptors, and employs hierarchical matching with adjustable thresholds. These parameter changes reduce computational complexity and processing time while preserving measurement precision
3Ease of operation
If traditional key frame addition based on distance is used, then ease of operation is improved, but reliability deteriorates due to rotation and viewpoint changes
Solution Approach 1:
The patent changes the key frame addition criterion from distance-based to descriptor-similarity-based. By using ORB descriptor comparison with thresholding, the system becomes reliable under rotation and viewpoint changes while remaining operationally simple through automated threshold-based selection
Data Source
AI summary
A system for place recognition is described herein. The system for place recognition comprises a plurality of sensors, a memory, and a processor. The memory is to store instructions and is communicatively coupled to the plurality of sensors. The processor is communicatively coupled to the plurality of sensors and the memory. When the processor is to execute the instructions, the processor is to detect features in a current frame and extract descriptors of the features of the current frame. The processor is also to generate a vocabulary tree using the descriptors and determine candidate key frames based on the vocabulary tree and detected features. The processor also is to perform place recognition via a first stage matching and a second stage matching.