A lightweight LBM detection transformer with multi-scale feature fusion for citrus-picking robots
Ke Gao, Baijing Wu, Jiren Gu, Guanghui Yan · 2026
Read A lightweight LBM detection transformer with multi-scale feature fusion for citrus-picking robots by Ke Gao, Baijing Wu, Jiren Gu, Guanghui Yan (2026) 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/PMC13311029.txt.
Description
Ke Gao and colleagues design a light vision model for citrus picking robots in a 2026 study. Orchard scenes hide small fruit in clutter, which causes missed picks and false alarms. The team built a detection transformer that fuses traits across scales while staying light for field robots.
A large kernel attention block pulls global, local, and texture cues at once. A linear attention encoder sharpens fruit focus against leaf noise, and a two way pyramid neck joins deep and shallow traits to recover hidden fruit. Against the base model, the design cut millions of parameters and lifted accuracy scores by clear margins on orchard tests.
Catalog record
| CREATOR | Ke Gao, Baijing Wu, Jiren Gu, Guanghui Yan |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 55645 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |