Ranging method and apparatus, robot, and storage medium
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cleaning robots face challenges with high-cost and large-volume laser radars, and complex camera algorithms leading to inaccurate obstacle detection, which are not suitable for miniaturized sweeping robots.
Innovation Solution
A ranging method using a first and second image collection apparatus to determine initial parallax and actual parallax distance, calculating depth information based on binocular measurement to accurately measure object distance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If laser radars are used to identify obstacles and calculate distances, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent uses image collection apparatus (cameras) to capture visual information of obstacles and their surroundings, creating a visual copy of the scene. By analyzing the position of obstacle images and ground images in the captured frames, the system calculates distance without requiring complex laser radar hardware. This copying approach replaces expensive sensing devices with simpler visual sensors while maintaining measurement capability.
Solution Approach 2:
The patent replaces the mechanical laser radar system with an optical-imaging-based system. Instead of using laser beams and time-of-flight measurements, the system uses image collection apparatus to capture images and employs geometric relationships (parallax) between obstacle images and ground images to calculate distances. This substitution reduces device complexity while achieving the same ranging function.
2Device complexity
If cameras are used to identify obstacles, then device complexity is reduced, but measurement precision deteriorates due to complex algorithms and low calculation accuracy
Solution Approach 1:
The patent introduces ground images as an intermediary element to improve measurement precision. By capturing both obstacle images and ground images in the same frame and analyzing their relative positions, the system uses the ground as a reference plane to accurately calculate distances. This intermediary approach provides geometric constraints that enhance calculation accuracy without adding device complexity.
Solution Approach 2:
The patent changes the approach from using complex image processing algorithms to using geometric parameter relationships. By focusing on the positional parameters of obstacle images and ground images within the captured frame and applying parallax principles, the system achieves accurate distance calculation through simple geometric relationships rather than complex computational algorithms.
3Device complexity
If monocular image analysis is used to calculate depth distance, then device complexity is reduced, but measurement precision deteriorates
Solution Approach 1:
The patent transitions from monocular (single-viewpoint) image analysis to a binocular-like approach by capturing both obstacle images and ground images. By analyzing the vertical distance between these two image elements in the captured frame and applying parallax principles, the system creates a second dimension of measurement information. This dimensional enhancement allows accurate depth calculation using simple image processing without requiring multiple cameras or complex hardware.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Accurately calculates object distance using binocular ranging, correcting monocular depth errors, and enhancing obstacle detection precision in miniaturized robots.
Implementation Method 1
a first image collected by the first image collection apparatus... a second image collected by the second image collection apparatus
Implementation Method 2
determining initial parallax based on the first distance... determining an actual parallax distance between the first image collection apparatus and the second image collection apparatus by using a location of the second target point
Data Source
AI summary
A ranging method includes determining a first distance of the object to be measured relative to the self-mobile robot, where content in the first image comprises at least the object and a surface on which the object is located; selecting a point located on the object from the first image as a 5 reference point; determining initial parallax based on the first distance; and determining a region of interest from a second image collected by a second image collection apparatus, and determining a location of the reference point from the region of interest as a first target point; determining a second target point from the region of interest based on the first target point, and determining an actual parallax distance between the first image collection apparatus and the second image collection apparatus; and calculating depth information of the object based on the actual parallax distance.