Dataset Article | Open Access

Volume 2025 |Article ID 100084 | https://doi.org/10.1016/j.plaphe.2025.100084

The Global Wheat Full Semantic Organ Segmentation (GWFSS) dataset

Zijian Wang,^1,16 Radek Zenkl,^6,16 Latifa Greche,^11,16 Benoit De Solan,¹ Lucas Bernigaud Samatan,^1,17 Safaa Ouahid,^2,15 Andrea Visioni,² Carlos A. Robles-Zazueta,^3,18 Francisco Pinto,^3,19 Ivan Perez-Olivera,^3,20 Matthew P. Reynolds,³ Chen Zhu,⁴ Shouyang Liu,⁴ Marie-Pia D'argaignon,^5,21 Raul Lopez-Lozano,⁵ Marie Weiss,⁵ Afef Marzougui,⁶ Lukas Roth,⁶ Sebastien Dandrifosse,^7,22 Alexis Carlier,^7,23 Benjamin Dumont,⁷ Benoît Mercatoris,⁷ Javier Fernandez,⁸ Scott Chapman,⁸ Keyhan Najafian,⁹ Ian Stavness,⁹ Haozhou Wang,¹⁰ Wei Guo,¹⁰ Nicolas Virlet,¹¹ Malcolm J. Hawkesford,¹¹ Zhi Chen,¹² Etienne David,¹³ Joss Gillet,¹⁴ Kamran Irfan,¹⁴ Alexis Comar,¹⁴ and Andreas Hund ⁶

¹Arvalis, France
²International Center for Agricultural Research in the Dry Areas, Rabat, Morocco
³Global Wheat Program, International Maize and Wheat Improvement Center, C.P. 56237, El Batan, Texcoco, Mexico
⁴Nanjing Agricultural University, China
⁵EMMAH, UMR1114, LPA CAPTE, INRAE, Avignon, France
⁶Department of Environmental Systems Science, ETH Zurich, 8092, Zurich, Switzerland
⁷Plant Sciences & Biosystems Dynamics and Exchanges, TERRA Teaching and Research Centre, Gembloux Agro-Bio Tech, University of Liege, Gembloux, Belgium
⁸School of Agriculture and Food Sustainability, The University of Queensland, Brisbane, Australia
⁹Department of Computer Science, University of Saskatchewan, Saskatoon, Canada
¹⁰Graduate School of Agricultural and Life Sciences, The University of Tokyo, Tokyo, Japan
¹¹Sustainable Soils and Crops, Rothamsted Research, West Common, Harpenden, AL5 2JQ, UK
¹²School of Electrical Engineering and Computer Science, The University of Queensland, Brisbane, Australia
¹³neoBloom, Munstermannskamp 1, 21335 Lüneburg, Germany
¹⁴HIPHEN SAS, 120 rue Jean Dausset, 84140 Avignon, France
¹⁵Programa de Doctorado de Ingeniería Agraria, Alimentaria, Forestal y del Desarrollo Rural Sostenible, Universidad de Cordoba, Cordoba, Spain
¹⁶These authors contributed equally to this work
¹⁷Current address: University of Toulouse, INRAE, UMR DYNAFOR, 31326 Castanet-Tolosan, France.
¹⁸Current address: Department of Plant Breeding, Hochschule Geisenheim University, 65366 Geisenheim, Germany.
¹⁹Current address: Centre for Crop Systems Analysis, Wageningen University & Research, Bornsesteeg 48, Building 109, 6708 PE Wageningen, The Netherlands.
²⁰Current address: Department of Agricultural and Biosystems Engineering, South Dakota State University, P. O. Box 57007, Brookings, SD, USA.
²¹Current address: Limagrain Chappes - Centre de Recherche; 28 Rte d’Ennezat 63720 Clermont-Ferrant.
²²Current address: Walloon Agricultural Research Centre, Gembloux, Belgium.
²³Current address: Osiris Agriculture, France.

Abstract

Computer vision is increasingly used in farmers' fields and agricultural experiments to quantify important traits. Imaging setups with a sub-millimeter ground sampling distance enable the detection and tracking of plant features, including size, shape, and colour. Although today's AI-driven foundation models segment almost any object in an image, they still fail for complex plant canopies. To improve model performance, the global wheat dataset consortium assembled a diverse set of images from experiments around the globe. After the head detection dataset (GWHD), the new dataset targets a full semantic segmentation (GWFSS) of organs (leaves, stems and spikes) covering all developmental stages. Images were collected by 11 institutions using a wide range of imaging setups. Two datasets are provided: i) a set of 1096 diverse images in which all organs were labelled at the pixel level, and (ii) a dataset of 52,078 images without annotations available for additional training. The labelled set was used to train segmentation models based on DeepLabV3Plus and Segformer. Our Segformer model performed slightly better than DeepLabV3Plus with a mIOU for leaves and spikes of ca. 90 %. However, the precision for stems with 54 % was rather lower. The major advantages over published models are: i) the exclusion of weeds from the wheat canopy, ii) the detection of all wheat features including necrotic and senescent tissues and its separation from crop residues. This facilitates further development in classifying healthy vs. unhealthy tissue to address the increasing need for accurate quantification of senescence and diseases in wheat canopies.