SAAIL Lab Publications
Agronomy · 2021 21+ Citations Prior Institutional Research · Pre-IITMZ (2021)

Detecting the Early Flowering Stage of Tea Chrysanthemum Using the F-YOLO Model

Qi Chao 1 , Innocent Nyalala 1 , Kunjie Chen 1
1College of Engineering, Nanjing Agricultural University, Nanjing, China
Agronomy (MDPI), Vol. 11(5), 834 (Apr 2021)

Summary: Introduces F-YOLO for accurate, lightweight detection of early flowering stages in field tea chrysanthemum images.

Abstract

Proposes the F-YOLO object detector tailored for detecting small, occluded early-stage tea chrysanthemum blossoms under complex outdoor lighting, enabling automated harvest timing.

Keywords: F-YOLOobject detectiontea chrysanthemumAgronomyprecision agriculture

Key Findings & Contributions

F-YOLO
Lightweight Detector
Single mobile GPU deployable
CSP Backbones
CSPDense + CSPResNeXt
Dual gradient-flow design
Multi-Scale
Fusion Detection Head
Handles occlusion & overlap
21+ Cites
Agronomy MDPI
Vol. 11(5), Article 834
  1. Combined cutout and mosaic input augmentation with a hybrid CSPDenseNet backbone and CSPResNeXt neck to detect early-stage tea chrysanthemum blossoms under outdoor occlusion.
  2. Fused multi-scale detection layers to maintain gradient diversity on dense small buds under shifting solar illumination.
  3. Demonstrated lightweight real-time inference feasible on a single mobile GPU for selective robotic chrysanthemum harvesting.

Benchmark Results

F-YOLO Architectural Modules for Selective Blossom Harvesting (Agronomy 2021)

Network ComponentModule ArchitectureAgronomic Purpose
Input StageCutout + Mosaic slicingRobustness to dense cluster overlapping
BackboneCSPDenseNet + feature fusionMaximizes gradient flow for small early buds
Neck & HeadCSPResNeXt + multi-scale fusionTruncates redundancy for mobile GPU latency

BibTeX

@article{chao2021fyolo,
  title={Detecting the Early Flowering Stage of Tea Chrysanthemum Using the F-YOLO Model},
  author={Chao, Qi and Nyalala, Innocent and Chen, Kunjie},
  journal={Agronomy},
  volume={11},
  number={5},
  pages={834},
  year={2021},
  doi={10.3390/agronomy11050834}
}
APA Citation

Chao, Q., Nyalala, I., & Chen, K. (2021). Detecting the Early Flowering Stage of Tea Chrysanthemum Using the F-YOLO Model. Agronomy, 11(5), 834. https://doi.org/10.3390/agronomy11050834