Assessing the capability of YOLO- and transformer-based object detectors for real-time weed detection

dc.contributor.authorAllmendinger, Alicia
dc.contributor.authorSaltık, Ahmet Oğuz
dc.contributor.authorPeteinatos, Gerassimos G.
dc.contributor.authorStein, Anthony
dc.contributor.authorGerhards, Roland
dc.date.accessioned2025-10-01T11:24:59Z
dc.date.available2025-10-01T11:24:59Z
dc.date.issued2025
dc.date.updated2025-07-03T12:50:37Z
dc.description.abstractSpot spraying represents an efficient and sustainable method for reducing herbicide use in agriculture. Reliable differentiation between crops and weeds, including species-level classification, is essential for real-time application. This study compares state-of-the-art object detection models-YOLOv8, YOLOv9, YOLOv10, and RT-DETR-using 5611 images from 16 plant species. Two datasets were created, dataset 1 with training all 16 species individually and dataset 2 with grouping weeds into monocotyledonous weeds, dicotyledonous weeds, and three chosen crops. Results indicate that all models perform similarly, but YOLOv9s and YOLOv9e, exhibit strong recall (66.58 % and 72.36 %) and mAP50 (73.52 % and 79.86 %), and mAP50-95 (43.82 % and 47.00 %) in dataset 2. RT-DETR-l, excels in precision reaching 82.44 % (dataset 1) and 81.46 % (dataset 2) making it ideal for minimizing false positives. In dataset 2, YOLOv9c attains a precision of 84.76% for dicots and 78.22% recall for Zea mays L.. Inference times highlight smaller YOLO models (YOLOv8n, YOLOv9t, and YOLOv10n) as the fastest, reaching 7.64 ms (dataset 1) on an NVIDIA GeForce RTX 4090 GPU, with CPU inference times increasing significantly. These findings emphasize the trade-off between model size, accuracy, and hardware suitability for real-time agricultural applications.en
dc.description.sponsorshipOpen Access funding enabled and organized by Projekt DEAL.
dc.description.sponsorshipUniversität Hohenheim (3153)
dc.identifier.urihttps://doi.org/10.1007/s11119-025-10246-0
dc.identifier.urihttps://hohpublica.uni-hohenheim.de/handle/123456789/17907
dc.language.isoeng
dc.rights.licensecc_by
dc.subjectWeed control
dc.subjectDigital farming
dc.subjectComputer vision
dc.subjectDeep learning
dc.subjectSingle stage detector
dc.subjectYOLO
dc.subjectDetection transformer
dc.subject.ddc630
dc.titleAssessing the capability of YOLO- and transformer-based object detectors for real-time weed detectionen
dc.type.diniArticle
dcterms.bibliographicCitationPrecision agriculture, 26 (2025), 3, 52. https://doi.org/10.1007/s11119-025-10246-0. ISSN: 1573-1618
dcterms.bibliographicCitation.articlenumber52
dcterms.bibliographicCitation.issn1573-1618
dcterms.bibliographicCitation.issue3
dcterms.bibliographicCitation.journaltitlePrecision agriculture
dcterms.bibliographicCitation.volume26
local.export.bibtex@article{Allmendinger2025, doi = {10.1007/s11119-025-10246-0}, url = {https://hohpublica.uni-hohenheim.de/handle/123456789/17907}, author = {Allmendinger, Alicia and Saltık, Ahmet Oğuz and Peteinatos, Gerassimos G. et al.}, title = {Assessing the capability of YOLO- and transformer-based object detectors for real-time weed detection}, journal = {Precision agriculture}, year = {2025}, volume = {26}, number = {3}, }
local.subject.sdg2
local.subject.sdg9
local.subject.sdg12
local.title.fullAssessing the capability of YOLO- and transformer-based object detectors for real-time weed detection

Files

Original bundle

Now showing 1 - 2 of 2
Loading...
Thumbnail Image
Name:
11119_2025_Article_10246.pdf
Size:
2.45 MB
Format:
Adobe Portable Document Format
Loading...
Thumbnail Image
Name:
11119_2025_10246_MOESM1_ESM.pdf
Size:
15.28 MB
Format:
Adobe Portable Document Format

License bundle

Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
7.85 KB
Format:
Item-specific license agreed to upon submission
Description: