Ground Texture Image Registration for SLAM Error Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing SLAM navigation methods face challenges in precision due to complex feature positions in images, leading to lower reliability in both complex and simple scenarios, affecting the accuracy of robot navigation.
Innovation Solution
A navigation method based on ground texture images that performs transform domain image registration to determine poses, inserts key-frame images into a map, and performs loop closure detection to correct accumulated errors, enhancing navigation accuracy in complex environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If feature points are detected using ORB algorithm for SLAM navigation, then the robot can map the environment and localize itself, but the precision of navigation is reduced in complex scenarios due to highly complex feature positions and difficulty in feature selection and extraction
Solution Approach 1:
The patent changes the parameter domain from spatial domain to transform domain (frequency domain) for image registration. By performing image registration in the transform domain, the system overcomes the limitations of feature-based methods in complex scenarios, achieving both high reliability and precision simultaneously.
Solution Approach 2:
The patent replaces the mechanical feature point detection and matching system with a transform domain-based image registration system. This substitution eliminates the need for complex feature selection and extraction, directly processing images in the transform domain to achieve accurate pose estimation.
2Measurement precision
If feature points are detected using ORB algorithm for SLAM navigation, then the robot can perform navigation in simple scenarios, but the reliability is lowered due to fewer feature points available
Solution Approach 1:
The patent transforms the image registration problem from spatial domain to transform domain, enabling reliable navigation in simple scenarios where feature points are scarce. The transform domain approach does not depend on the quantity of feature points, thus maintaining both precision and reliability.
3Measurement precision
If transform domain based image registration is performed to determine pose, then navigation accuracy is improved, but accumulated errors still occur in the map requiring correction through loop closure detection
Solution Approach 1:
The patent implements loop closure detection as a feedback mechanism to detect and correct accumulated errors in the map. When the robot returns to a previously visited location, the system detects the loop closure and adjusts the map accordingly, ensuring long-term map accuracy while maintaining high navigation precision.
Data Source
AI summary
A navigation method based on ground texture images, an electronic device and storage medium. The method includes: performing transform domain based image registration on an acquired image of a current frame and an image of a previous frame, and determining a first pose of the image of the current frame; determining whether the image of the current frame meets a preset condition, and if so, inserting the image of the current frame as the key-frame image into a map, and performing loop closure detection and determining a loop key-frame image; performing transform domain based image registration on the image of the current frame and the loop key-frame image, and determining a second pose of the image of the current frame; and determining an accumulated error according to the first pose and the second pose, and correcting the map according to the accumulated error.