Optimized YOLO architectures for efficient Kiwi detection in precision agriculture on embedded systems

Berat Karacaoglu, Muhammet Emin Sahin · 2025

Read Optimized YOLO architectures for efficient Kiwi detection in precision agriculture on embedded systems by Berat Karacaoglu, Muhammet Emin Sahin (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-12749734.txt.

Description

Counting kiwifruit on the vine, hidden under leaves, is a job for cameras that must run on small field computers. Karacaoglu and Sahin benchmark five generations of YOLO detectors, from the eighth to the eleventh, each in five sizes, on nearly five thousand labeled field photos. They then tune the small models' settings in a structured search and test everything again on a Jetson TX2 edge board.

The tuned YOLOv11s gave the best balance, scoring 0.956 on the standard accuracy measure with a few seconds per image on the small board. Bigger models scored a hair higher but ran too slowly for the field. The authors offer their pipeline as a template for other fruit tasks.

Catalog record

CREATORBerat Karacaoglu, Muhammet Emin Sahin
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE81 KB
VERSION1.0
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC12749734

Browse the full library · About the Tower of Babel