Robot Monocular Positioning Using Marker Pattern Pose Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Monocular visual positioning for micro-robots is computationally intensive due to the large overhead in calculation, which is a challenge given their small size and processing capabilities.
Innovation Solution
A positioning method that involves a camera on the robot taking photos of a marker pattern area, identifying marker points, calculating coordinate data, and determining the robot's pose using optical principles and camera parameters, reducing the computational load by recognizing marker points and converting coordinates efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If monocular visual positioning technology is used for micro-robots, then positioning capability is achieved, but computational load becomes excessive for the processor
Solution Approach 1:
The patent introduces a marker pattern as an intermediary object between the robot and the positioning system. The marker contains encoded position information that the robot's camera can directly read, eliminating the need for complex monocular visual positioning calculations. This mediator transfers the positioning function from computational processing to optical recognition, significantly reducing the processor's computational load while maintaining positioning accuracy.
2Measurement precision
If complex calculation methods are used for positioning, then positioning accuracy is improved, but processing time increases
Solution Approach 1:
The marker pattern is pre-designed with encoded position information that directly represents the robot's location. This preliminary encoding of position data in the marker allows the robot to determine its position through simple image recognition and code decoding, rather than performing complex real-time calculations. The positioning information is prepared in advance in the marker's pattern, enabling rapid retrieval and processing.
Data Source
AI summary
The present disclosure provides a positioning method and a robot with the same. The method is applied to a robot having a camera, the method includes: taking a photo of a predetermined area, wherein the predetermined area comprises a marker pattern area; identifying at least two marker points of the marker pattern area and obtaining first coordinate data of each marker point; calculating second coordinate data of the robot in a world coordinate based on parameters of the camera and the first coordinate data of the marker points; and calculating angle data of the robot in the world coordinate based on the first coordinate data of the marker points and the second coordinate data of the robot in the world coordinate. Through the above-mentioned positioning method, the robot can quickly perform monocular visual positioning and reduce the calculation amount of positioning.