WeedNet: a foundation model-based global-to-local AI approach for weed species identification and classification
7 affiliations
- 1Department of Agronomy, Iowa State University, Ames, IA, USA
- 2Department of Mechanical Engineering, Iowa State University, Ames, IA, USA
- 3Department of Computer Science, Iowa State University, Ames, IA, USA
- 4Data Science Institute, University of Arizona, Tucson, AZ, USA
- 5Department of Plant Pathology, Entomology and Microbiology, Iowa State University, Ames, IA, USA
- 6Department of Agronomy, Iowa State University, Ames, IA, USA. [email protected]
- 7Department of Mechanical Engineering, Iowa State University, Ames, IA, USA. [email protected]
Abstract
Early weed identification is crucial for effective management and control, and researchers, agronomists, and technology developers are increasingly interested in automating this process using computer vision and artificial intelligence; however, limited expert-verified data and variable morphological features have hindered the development of AI-based weed identification models. To address these issues, we present WeedNet, a global-scale weed identification model that can recognize an extensive set of weed species. WeedNet is an end-to-end weed identification pipeline that uses self-supervised learning, fine-tuning, and enhanced trustworthiness strategies. WeedNet achieved 91.02% accuracy across 1,593 weed species, with 41% of species achieving 100% accuracy. Using a fine-tuning approach, the local Iowa WeedNet model achieved 97.38% overall accuracy for 84 Iowa weeds. Testing across intra-species dissimilarity and inter-species similarity suggests that diversity in the images collected, spanning all the growth stages and distinguishable plant characteristics, is crucial in driving model performance. The global WeedNet model serves as a foundation, and the global-to-local strategy enables targeted fine-tuning to improve performance in regional weed communities. Additional validation of drone- and ground-rover-based images highlights WeedNet's potential for integration into robotic platforms. Furthermore, integration with artificial intelligence for conversational use provides intelligent agricultural and ecological conservation consulting tools for farmers, researchers, and government agencies across diverse landscapes.
Related
- Constructing a lower-bound estimate of the global number of insect species on a hyperdiverse empirical foundation
- Accurate data as the foundation of equitable cancer control: Lessons from global cancer statistics 2024 and the complexity of ovarian cancer
- Creator 1.0: Phylogeny-Guided Stagewise Pretraining for Plant Genome Foundation Models across 1,000 Species
- LucaCell: a sequence-centric foundation model for cross-species single-cell analysis
- Bioactivity foundation model for hit identification and hit-to-lead optimization from sequence data
BibTeX
@article{shen2026weednet,
title = {{WeedNet: a foundation model-based global-to-local AI approach for weed species identification and classification}},
author = {Yanben Shen and Timilehin T. Ayanlade and Venkata Naresh Boddepalli and Mojdeh Saadati and Ashlyn M. Rairdin and Zi K. Deng and Muhammad Arshad and Shivani Chiranjeevi and Aditya Balu and Daren Mueller and Asheesh K. Singh and Wesley Jay Everman and Nirav Merchant and Baskar Ganapathysubramanian and Meaghan J. B. Anderson and Soumik Sarkar and Arti Singh},
journal = {Nature Communications},
year = {2026},
month = sep,
doi = {10.1038/s41467-026-78044-4},
url = {https://www.nature.com/articles/s41467-026-78044-4},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - JOUR
TI - WeedNet: a foundation model-based global-to-local AI approach for weed species identification and classification
AU - Yanben Shen
AU - Timilehin T. Ayanlade
AU - Venkata Naresh Boddepalli
AU - Mojdeh Saadati
AU - Ashlyn M. Rairdin
AU - Zi K. Deng
AU - Muhammad Arshad
AU - Shivani Chiranjeevi
AU - Aditya Balu
AU - Daren Mueller
AU - Asheesh K. Singh
AU - Wesley Jay Everman
AU - Nirav Merchant
AU - Baskar Ganapathysubramanian
AU - Meaghan J. B. Anderson
AU - Soumik Sarkar
AU - Arti Singh
PY - 2026
DA - 2026/09/30
JO - Nature Communications
DO - 10.1038/s41467-026-78044-4
UR - https://www.nature.com/articles/s41467-026-78044-4
AB - Early weed identification is crucial for effective management and control, and researchers, agronomists, and technology developers are increasingly interested in automating this process using computer vision and artificial intelligence; however, limited expert-verified data and variable morphological features have hindered the development of AI-based weed identification models. To address these issues, we present WeedNet, a global-scale weed identification model that can recognize an extensive set of weed species. WeedNet is an end-to-end weed identification pipeline that uses self-supervised learning, fine-tuning, and enhanced trustworthiness strategies. WeedNet achieved 91.02% accuracy across 1,593 weed species, with 41% of species achieving 100% accuracy. Using a fine-tuning approach, the local Iowa WeedNet model achieved 97.38% overall accuracy for 84 Iowa weeds. Testing across intra-species dissimilarity and inter-species similarity suggests that diversity in the images collected, spanning all the growth stages and distinguishable plant characteristics, is crucial in driving model performance. The global WeedNet model serves as a foundation, and the global-to-local strategy enables targeted fine-tuning to improve performance in regional weed communities. Additional validation of drone- and ground-rover-based images highlights WeedNet's potential for integration into robotic platforms. Furthermore, integration with artificial intelligence for conversational use provides intelligent agricultural and ecological conservation consulting tools for farmers, researchers, and government agencies across diverse landscapes.
SN - 2041-1723
ER -