Virtual Plant Model Generation via Skeleton Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image recognition systems face difficulties in identifying individual plants in close proximity due to overlapping growths, often mischaracterizing plant features or miscounting the number of plants, which hampers precision in agricultural applications.
Innovation Solution
A method for generating virtual plant models by detecting unique plant features like stalk bases from images, generating point clouds, and classifying skeleton segments to create detailed plant representations, enabling accurate plant parameter estimation and computation speed improvements through downsampled point clouds and neural network classification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If plants are planted close in proximity to maximize yield, then land usage efficiency is improved, but plant identification accuracy deteriorates due to overlapping growths
Solution Approach 1:
The patent segments plant identification into multiple stages: detecting unique plant features (stalk bases), generating point clouds, creating skeleton segments, and classifying segments as stalk bases or leaves. This segmentation allows the system to distinguish individual plants even when their foliage overlaps, resolving the contradiction between high plant density and identification accuracy.
Solution Approach 2:
The patent introduces intermediate representations (point clouds and skeleton segments) as mediators between raw images and final plant identification. These intermediates capture spatial relationships and structural information that enable accurate plant separation and identification even in dense plantings where direct image analysis fails.
2Device complexity
If existing image recognition systems are used for plant identification, then system simplicity is maintained, but plant feature characterization accuracy deteriorates
Solution Approach 1:
The patent segments the plant identification process into distinct computational stages: image input, point cloud generation, skeleton segmentation, feature classification, and virtual model generation. This structured segmentation improves accuracy by processing information in optimized stages rather than attempting single-step analysis.
Solution Approach 2:
The patent transitions from 2D image analysis to 3D point cloud representations and skeletal structures. This dimensional transformation enables the system to capture spatial relationships and plant architecture that are lost in traditional 2D image recognition, significantly improving feature characterization accuracy.
3Measurement precision
If detailed plant modeling is performed, then plant parameter estimation accuracy is improved, but computation time increases
Solution Approach 1:
The patent segments the computational process into efficient stages: generating point clouds from images, creating skeleton segments, downsampled point cloud processing, and neural network classification. This segmentation enables detailed modeling while managing computation time through optimized processing stages.
Solution Approach 2:
The patent uses downsampled point clouds to represent plant structures at reduced resolution for classification purposes. This partial representation approach maintains sufficient accuracy for plant parameter estimation while significantly reducing computational requirements compared to processing full-resolution detailed models.
Data Source
AI summary
A technique for generating virtual models of plants in a field is described. Generally, this includes recording images of plants in-situ; generating point clouds from the images; generating skeleton segments from the point cloud; classifying a subset of skeleton segments as unique plant features using the images; and growing plant skeletons from skeleton segments classified as unique plant feature. The technique may be used to generate a virtual model of a single, real plant, a portion of a real plant field, and/or the entirety of the real plant field. The virtual model can be analyzed to determine or estimate a variety of individual plant or plant population parameters, which in turn can be used to identify potential treatments or thinning practices, or predict future values for yield, plant uniformity, or any other parameter can be determined from the projected results based on the virtual model.