EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

DSC-DeepLabv3+: a lightweight semantic segmentation model for weed identification in maize fields

Haitao Fu, Xiaoyao Li, Li Zhu, Xin Pan, Tuo Wu, Wen Li, Yuxuan Feng

Frontiers in Plant Science · 03.09.2025 · 16

DOI: 10.3389/fpls.2025.1647736 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Frontiers in Plant Science, online erschienen am 03.09.2025. Verfasst von Haitao Fu und 6 weiteren Autor:innen. Beteiligte Einrichtung: College of Information Technology, Jilin Agricultural University, Changchun, China. Die Arbeit wurde bisher 14-mal zitiert (Crossref) und nennt 43 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Introduction

Weeds compete with crops for water, nutrients, and light, negatively impacting maize yield and quality. To enhance weed identification accuracy and meet the requirements of precision agriculture, we propose a lightweight semantic segmentation model named DSC-DeepLabv3+.

Methods

MobileNetV2 is adopted as the backbone, and standard convolutions in atrous spatial pyramid pooling (ASPP) and decoder modules are replaced with depthwise separable dilated convolutions (DSDConv), significantly reducing model complexity and improving segmentation efficiency. To capture rich contextual information, strip pooling is incorporated into the ASPP module, forming the strip pooling–atrous spatial pyramid pooling (S-ASPP) structure. In addition, a convolutional block attention module (CBAM) is introduced to refine feature representations, and multi-scale features are further fused using the CBAM–Cascade Feature Fusion (C-CFF) module to improve semantic understanding.

Results

Experimental results show that the proposed model reduces the number of parameters from 54.714M to 2.89M and decreases the computational cost from 167.139 GFLOPs to 15.326 GFLOPs, while achieving an inference speed of 42.89 FPS and a mean Intersection over Union (mIoU) of 85.57%.

Discussion

These results demonstrate that DSC-DeepLabv3+ strikes an effective balance between accuracy and efficiency, outperforming several classical lightweight models, making it a promising solution for accurate and efficient weed segmentation in agricultural applications.

Schlagwörter

Stichwörter der Autor:innen

DeepLabV3+attention mechanismsfeature fusionlightweight semantic segmentationweed recognition

Autor:innen und Institutionen

  • College of Information Technology, Jilin Agricultural University, Changchun, China.
  • BaichengAgricultural and Rural Information Center, Baicheng, China.

Zitieren

APA

Fu, H., Li, X., Zhu, L., Pan, X., Wu, T., Li, W., & Feng, Y. (2025). DSC-DeepLabv3+: a lightweight semantic segmentation model for weed identification in maize fields. Frontiers in Plant Science, 16. https://doi.org/10.3389/fpls.2025.1647736

Vancouver

Fu H, Li X, Zhu L, Pan X, Wu T, Li W, et al. DSC-DeepLabv3+: a lightweight semantic segmentation model for weed identification in maize fields. Frontiers in Plant Science. 2025;16. doi:10.3389/fpls.2025.1647736

BibTeX
@article{Fu2025,
  author = {Fu, Haitao and Li, Xiaoyao and Zhu, Li and Pan, Xin and Wu, Tuo and Li, Wen and Feng, Yuxuan},
  title = {DSC-DeepLabv3+: a lightweight semantic segmentation model for weed identification in maize fields},
  journal = {Frontiers in Plant Science},
  year = {2025},
  volume = {16},
  doi = {10.3389/fpls.2025.1647736}
}
Kostenlos starten →