Robot Map Building Using Passive Vision Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current robot positioning and map building methods rely on active signal emission, such as lasers or sonar, which are costly, power-intensive, and limited in environmental applicability, restricting the application of artificial intelligence in robots.
Innovation Solution
A map building method using a passive vision sensor that captures real-time scene images to build and update maps, including a first map for structural data and a second map for scene features, allowing for accurate robot positioning and environmental awareness without active signal emission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If active signal emission (laser, sonar) is used for robot positioning and map building, then positioning accuracy and environmental detection capability are improved, but power consumption increases and device cost increases
Solution Approach 1:
The patent replaces active signal emission systems (laser rangefinders, sonar) with a passive vision-based system using cameras and image processing. The robot captures images of the environment and uses computer vision algorithms to extract spatial information, feature points, and depth data, thereby substituting mechanical/optical active sensing with passive optical sensing and computational processing.
Solution Approach 2:
The patent creates a visual map (copy of the environment) by capturing images and extracting feature points, which then serves as a reference for positioning. Instead of directly measuring distances with active signals, the system creates a visual representation of the environment and uses image matching and feature correspondence to determine robot position and orientation.
2Measurement precision
If active signal emission (laser, sonar) is used for robot positioning and map building, then positioning accuracy and environmental detection capability are improved, but device cost increases
Solution Approach 1:
The patent uses inexpensive camera modules instead of expensive laser rangefinders or sonar systems. Cameras are mass-produced at low cost compared to active sensing devices, allowing the robot to achieve positioning and mapping capabilities with significantly reduced hardware cost while maintaining adequate performance through software-based image processing.
Solution Approach 2:
The patent replaces costly mechanical/optical active sensing systems with a vision-based system using standard camera hardware and computational algorithms, thereby reducing device cost while maintaining positioning and mapping functionality.
3Measurement precision
If active signal emission is used for robot positioning, then positioning capability is improved, but adaptability to various environments deteriorates
Solution Approach 1:
The patent changes the sensing parameter from active signal emission to passive light reception, allowing the system to adapt to various environments including dark areas where laser or sonar may not perform well. The vision system can process images under different lighting conditions and adapt to diverse environmental contexts through image processing algorithms.
Solution Approach 2:
The patent makes the positioning system universal by using a camera that can capture a wide range of environmental features (textures, colors, shapes, patterns) that serve as positioning references. The vision-based approach can work in various environments (indoor, outdoor, structured, unstructured) where active signal systems may fail or require specific conditions.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present disclosure provides a robot, a map building method, a positioning method and a storage medium. A map building method for a robot comprises: a robot traversing a work area according to a predetermined rule, and building an initial map according to a scene image captured by the robot in real time in the process of traverse, wherein the initial map comprises a first map including a mapping of the work area and a map coordinate system, and a second map including a scene feature extracted based on the captured scene image, a geometric quantity of the scene feature and a scene image extracting the scene feature, which are stored in association with the pose of the robot when capturing the scene image, the pose including a coordinate and an orientation of the robot in the first map, and the scene feature including a feature of a feature object in the scene and/or a feature of the scene image.