Weeding Robot Obstacle Recognition Without Boundary Wires
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing weeding robots require manual calibration using boundary wires, consuming significant manpower and material resources and limiting the shape of the weeding region due to wire-burying constraints.
Innovation Solution
An obstacle recognition method and apparatus that determines a candidate obstacle region in a weeding region image using color, contour, and chrominance information to identify obstacles without boundary wires, enhancing recognition efficiency and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If boundary wires are buried to calibrate boundaries of a weeding region, then the weeding robot can identify the working area, but a lot of manpower and material resources are consumed and costs are increased
Solution Approach 1:
The patent extracts the boundary calibration function from the physical boundary wires and transfers it to natural feature points in the environment (such as building corners, road intersections, etc.). The robot uses visual recognition to identify these feature points and calculate its position and orientation relative to the weeding region, thereby eliminating the need for boundary wires and reducing material resources.
Solution Approach 2:
The patent replaces the mechanical boundary wire system with a visual recognition system. Instead of using physical wires that require manual installation and calibration, the robot uses cameras and image processing algorithms to detect natural feature points and determine its position, substituting mechanical calibration with optical detection and computational geometry.
2Measurement precision
If boundary wires are buried to calibrate boundaries of a weeding region, then the weeding robot can identify the working area, but a shape of the weeding region is limited to some extent
Solution Approach 1:
The patent transforms the static boundary wire system into a dynamic visual recognition system. The robot can adapt to various shapes of weeding regions by dynamically identifying different combinations of natural feature points and calculating its position and orientation accordingly. This allows the system to handle arbitrary polygonal shapes without being constrained by the geometric limitations of wire-based calibration.
Solution Approach 2:
The patent creates a universal positioning method that can work with any polygonal shape of weeding region. By using natural feature points that are universally present in urban environments (buildings, roads, etc.), the system can calibrate boundaries for rectangular, triangular, irregular, and other complex shapes, making the solution universally applicable rather than shape-specific.
3Productivity
If color information is used to determine candidate obstacle regions, then obstacle recognition efficiency is improved, but recognition accuracy may be affected by similar-colored objects
Solution Approach 1:
The patent segments the obstacle recognition process into multiple stages: first using color information for rapid candidate region identification, then applying contour analysis and chrominance-value feature extraction to verify and refine the identification. This multi-stage segmentation allows the system to maintain high efficiency while improving accuracy by progressively filtering out false positives.
Solution Approach 2:
The patent implements a feedback mechanism where the results from contour analysis and chrominance-value extraction are used to verify the candidate obstacle regions identified by color information. If the subsequent analysis does not confirm the presence of an obstacle, the system can reject the initial color-based identification, thereby correcting errors and improving overall recognition accuracy.
Data Source
AI summary
An obstacle recognition method includes determining a candidate obstacle region in a candidate weeding region image according to color information of the candidate weeding region image; obtaining contour information of the candidate obstacle region and chrominance information and value information of the candidate weeding region image; and determining, according to the contour information and the chrominance information, or the contour information, the chrominance information, and the value information, whether there is an obstacle in the candidate weeding region image. A related obstacle recognition apparatus, electronic device, computer-readable storage medium, and weeding robot are also disclosed.