Fruit Bunch Picking Robot Branch Recognition via Spatial Posture
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Fruit bunch picking robots face challenges in accurately recognizing the correct branch during fruit bunch picking due to complex agricultural environments and variations in tomato fruit bunch posture and shape, leading to picking failures.
Innovation Solution
A modeling method for a picking target of a fruit bunch picking robot that involves obtaining an image of the to-be-picked region, extracting image features of branches and fruit clusters using a multi-task perception network, determining the correct fruit cluster and connected branch through a subordinate decision model, and modeling key points of the fruit cluster and branch for accurate targeting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If the robot uses a visual system to recognize fruit bunches in complex agricultural environments, then the robot can perform automated picking, but the robot often cannot recognize the correct branch due to similar colors and sheltered positions of multiple tomato plants
Solution Approach 1:
The patent segments the branch recognition task into multiple stages: first identifying the fruit cluster, then determining its spatial posture, and finally locating the connected branch based on the fruit cluster's position and orientation. This multi-stage segmentation approach resolves the contradiction by breaking down the difficult single-step branch recognition into manageable steps, improving accuracy while maintaining automation.
Solution Approach 2:
The patent introduces spatial posture information as an additional dimension for branch recognition. Instead of relying solely on color and shape features in the image plane, the system uses three-dimensional spatial relationships (position, orientation, and posture of the fruit cluster) to infer the correct branch. This dimensional enhancement allows the robot to distinguish between similarly colored branches by analyzing their spatial configurations relative to the fruit cluster.
2Productivity
If the robot attempts to recognize multiple tomato plants simultaneously, then the robot can handle large-scale planting bases, but the robot cannot differentiate between plants due to their similar colors and sheltered positions
Solution Approach 1:
The patent applies preliminary action by first identifying and isolating the to-be-picked fruit cluster before attempting to recognize its associated branch. The system pre-processes the visual information to focus on the target fruit cluster, determines its spatial posture in advance, and then uses this information to guide the branch search. This preliminary focus on the specific fruit cluster improves reliability in plant differentiation while maintaining productivity in large-scale operations.
Solution Approach 2:
The patent applies local quality by concentrating the recognition effort on the local region around the identified fruit cluster rather than attempting to analyze all tomato plants simultaneously. The system uses the fruit cluster's spatial posture to define a local search area for the connected branch, filtering out unrelated branches from other plants. This localized approach improves differentiation accuracy while maintaining efficiency in large-scale harvesting.
3Device complexity
If the robot uses traditional image recognition methods, then the system complexity remains low, but the robot cannot accurately determine the spatial posture and connected branch of the fruit cluster
Solution Approach 1:
The patent introduces spatial posture information as an intermediary between fruit cluster identification and branch location. Instead of directly matching fruit clusters to branches using complex image recognition, the system first determines the spatial posture (position, orientation) of the fruit cluster, then uses this intermediate information to guide the branch search. This intermediary approach achieves high measurement precision while avoiding the need for extremely complex direct recognition systems.
Data Source
AI summary
Disclosed is a modeling method for a picking target of a fruit bunch picking robot, which relates to the technical field of general image data processing or generation. The modeling method includes: obtaining an image of a to-be-picked region of a picking robot, and extracting image features of each branch and fruit cluster in the image of the to-be-picked region through a multi-task perception network; determining a to-be-picked fruit cluster based on the image feature of the fruit cluster; inputting the image features of the branch and the fruit cluster into a subordinate decision model to determine a branch connected to the to-be-picked fruit cluster; and extracting key points of image features of the to-be-picked fruit cluster and the branch connected to the to-be-picked fruit cluster, and modeling the to-be-picked fruit cluster and the branch connected to the to-be-picked fruit cluster.