A scalable multi-modal learning fruit detection algorithm for dynamic environments
Liang Mao, Zihao Guo, Mingzhe Liu, Yue Li, Linlin Wang, Jie Li · 2025
Read A scalable multi-modal learning fruit detection algorithm for dynamic environments by Liang Mao, Zihao Guo, Mingzhe Liu, Yue Li, Linlin Wang, Jie Li (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-11841473.txt.
Description
Liang Mao and co-authors build a fruit-spotting system for orchards in a study published in 2024. Their target is litchi fruit in real groves, where bunches hide behind leaves and the fruits are small. The method, YOLOv5-Litchi, adapts a popular real-time detection network to that hard scene.
The team slimmed the network's neck, added detection heads until there were five, and swapped two heads for a design that catches small targets better. Two of the training losses were replaced with sharper versions to tighten the boxes and cut missed fruit. A sliding slice tactic then scans big images in tiles so tiny fruits are not lost in the shrink. The goal is a detector that robotic and survey systems can trust in messy, moving orchard light.
Catalog record
| CREATOR | Liang Mao, Zihao Guo, Mingzhe Liu, Yue Li, Linlin Wang, Jie Li |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 45 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |