Related Experiment Video
Updated: Apr 28, 2026
08:47
Author Spotlight: UAV Remote Sensing for Efficient Invasive Plant Biomass Estimation
Published on: February 9, 2024
2.0K
Deep learning for sorghum yield forecasting using uncrewed aerial systems and lab-derived imagery.
Md Abdullah Al Bari1,2, Aliva Bakshi3, Jahid Chowdhury Choton3
1Department of Agronomy, Kansas State University, Manhattan, KS, 66506, USA.
Plant Phenomics (Washington, D.C.)
|April 27, 2026
Summary
Machine learning and deep learning extract plant traits from images for accurate yield prediction. YOLO models excel at detecting sorghum panicles, driving yield forecasts with high precision.
Area of Science:
- Agricultural Science
- Computer Science
- Genetics
Background:
- Advancements in AI, GPUs, and open-source platforms facilitate Machine Learning (ML) and Deep Learning (DL) for rapid phenotypic feature extraction from imagery.
- Phenotypic digitization and yield forecasting are crucial for assessing genotypes and advancing cultivar development.
Purpose of the Study:
- To evaluate the efficacy of YOLO and Faster R-CNN models in extracting yield-predictive features from Unmanned Aerial System (UAS) imagery of sorghum.
- To compare the performance of different regression models (SVR, RFR, DTR) for yield prediction using extracted features.
- To identify key phenotypic drivers influencing sorghum yield prediction.
Main Methods:
- Conducted a field trial with 36 sorghum genotypes in a Randomized Complete Block Design (RCBD).
- Acquired high-resolution field images using a DJI M300 drone at nadir and oblique angles.
- Trained YOLO and Faster R-CNN (Detectron2) models for panicle detection and feature extraction from UAS and lab images.
- Utilized Support Vector Regression (SVR), Random Forest Regression (RFR), and Decision Tree Regression (DTR) for yield prediction.
Main Results:
- YOLO models achieved superior sorghum panicle detection (mAP@0.50: 0.92-0.98) compared to Faster R-CNN (0.61-0.89).
- Field panicle detection correlated strongly (0.86) with ground truth counts.
- Lab image analysis showed high correlations for panicle area (0.79) and seed count (0.94).
- Yield prediction models achieved correlation coefficients of 0.74 (SVR), 0.71 (RFR), and 0.78 (DTR).
- SHapley Additive exPlanation (SHAP) analysis identified panicle seed count as the primary yield driver.
Conclusions:
- YOLO models are highly suitable for extracting yield-predictive phenotypic traits from UAS imagery.
- Integrating image-derived features into ML regression models enhances sorghum yield prediction accuracy.
- This approach supports efficient cultivar development through precise yield forecasting.
Related Concept Videos
02:16Light Acquisition
8.0K
In order to produce glucose, plants need to capture sufficient light energy. Many modern plants have evolved leaves specialized for light acquisition. Leaves can be only millimeters in width or tens of meters wide, depending on the environment. Due to competition for sunlight, evolution has driven the evolution of increasingly larger leaves and taller plants, to avoid shading by their neighbors with contaminant elaboration of root architecture and mechanisms to transport water and nutrients.
8.0K
01:29Application of Linearization and Approximation
192
A drone flying through complex terrain often relies on more than one sensing method to estimate small changes in altitude. Along with direct measurements, air pressure provides a useful indirect indicator of vertical movement. Atmospheric pressure decreases as altitude increases, and this relationship is commonly described using an exponential model. Although accurate, converting pressure measurements into altitude values requires calculations that are too complex to perform repeatedly during...
192