UAV Phenotyping Data for Megathyrsus maximus Yield and Height Prediction
by Guilherme Francio Niederauer·Updated 1mo ago
16.8 MB1files
Available on 1 platform
Sign in to view source links and access this dataset
Description
A study by Guilherme Francio Niederauer, published on figshare in April 2026, optimized UAV-based phenotyping methods for a Megathyrsus maximus biparental population. It examines how ground sampling distance, environment, and harvest date affect the accuracy of RGB-derived digital traits in predicting yield and canopy height, applying machine learning and mixed model analyses.
Use Cases
Predicting green and dry matter yield based on pixel count and Haralick's entropy features.
Estimating canopy height using machine learning models trained on UAV-derived digital traits.
Evaluating genotype selection efficiency by comparing rankings from digital traits with conventional yield measurements.
Assessing the heritability and environmental susceptibility of various vegetative indices and canopy height.
Strengths
Demonstrates strong predictive power for yield with machine learning models achieving correlations greater than 0.80.
Reports high broad-sense heritability (0.7 to 0.87) for key yield traits, pixel count, and entropy.
Identifies an optimal ground sampling distance range of 0.5–1.0 cm for UAV image acquisition.
Shows 80% correspondence between top genotypes ranked by pixel count and conventional dry matter yield.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count is unknown, which may limit suitability assessment.
The dataset is contained within a 16.8 MB PDF file, which may require extraction of underlying data.
Provenance
Source
figshare
Collection Method
UAV-based image acquisition and analysis of a Megathyrsus maximus biparental population.
Time Range
Harvest dates were examined, but specific temporal coverage is not provided.
Freshness
Last updated 2026-04-22 05:46:04; freshness should be verified.
Geography
Environment 2 is mentioned, but specific geographic location is not provided.
Primary data is embedded within a PDF document; users may need to extract tabular or image data for analysis.