MSEA-Net: Multi-Scale and Edge-Aware Network for Weed Segmentation

Accurate weed segmentation in Unmanned Aerial Vehicle (UAV) imagery remains a significant challenge in precision agriculture due to environmental variability, weak contextual representation, and inaccurate boundary detection. To address these limitations, we propose the Multi-Scale and Edge-Aware Ne...

Full description

Saved in:
Bibliographic Details
Main Authors: Akram Syed, Baifan Chen, Adeel Ahmed Abbasi, Sharjeel Abid Butt, Xiaoqing Fang
Format: Article
Language:English
Published: MDPI AG 2025-04-01
Series:AgriEngineering
Subjects:
Online Access:https://www.mdpi.com/2624-7402/7/4/103
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850156110568226816
author Akram Syed
Baifan Chen
Adeel Ahmed Abbasi
Sharjeel Abid Butt
Xiaoqing Fang
author_facet Akram Syed
Baifan Chen
Adeel Ahmed Abbasi
Sharjeel Abid Butt
Xiaoqing Fang
author_sort Akram Syed
collection DOAJ
description Accurate weed segmentation in Unmanned Aerial Vehicle (UAV) imagery remains a significant challenge in precision agriculture due to environmental variability, weak contextual representation, and inaccurate boundary detection. To address these limitations, we propose the Multi-Scale and Edge-Aware Network (MSEA-Net), a lightweight and efficient deep learning framework designed to enhance segmentation accuracy while maintaining computational efficiency. Specifically, we introduce the Multi-Scale Spatial-Channel Attention (MSCA) module to recalibrate spatial and channel dependencies, improving local–global feature fusion while reducing redundant computations. Additionally, the Edge-Enhanced Bottleneck Attention (EEBA) module integrates Sobel-based edge detection to refine boundary delineation, ensuring sharper object separation in dense vegetation environments. Extensive evaluations on publicly available datasets demonstrate the effectiveness of MSEA-Net, achieving a mean Intersection over Union (IoU) of 87.42% on the Motion-Blurred UAV Images of Sorghum Fields dataset and 71.35% on the CoFly-WeedDB dataset, outperforming benchmark models. MSEA-Net also maintains a compact architecture with only 6.74 M parameters and a model size of 25.74 MB, making it suitable for UAV-based real-time weed segmentation. These results highlight the potential of MSEA-Net for improving automated weed detection in precision agriculture while ensuring computational efficiency for edge deployment.
format Article
id doaj-art-4e31238d40c5406785702863711cb641
institution OA Journals
issn 2624-7402
language English
publishDate 2025-04-01
publisher MDPI AG
record_format Article
series AgriEngineering
spelling doaj-art-4e31238d40c5406785702863711cb6412025-08-20T02:24:41ZengMDPI AGAgriEngineering2624-74022025-04-017410310.3390/agriengineering7040103MSEA-Net: Multi-Scale and Edge-Aware Network for Weed SegmentationAkram Syed0Baifan Chen1Adeel Ahmed Abbasi2Sharjeel Abid Butt3Xiaoqing Fang4School of Automation, Central South University, Changsha 410083, ChinaSchool of Automation, Central South University, Changsha 410083, ChinaSchool of Computer Science, Central South University, Changsha 410083, ChinaSchool of Automation, Central South University, Changsha 410083, ChinaSchool of Automation, Central South University, Changsha 410083, ChinaAccurate weed segmentation in Unmanned Aerial Vehicle (UAV) imagery remains a significant challenge in precision agriculture due to environmental variability, weak contextual representation, and inaccurate boundary detection. To address these limitations, we propose the Multi-Scale and Edge-Aware Network (MSEA-Net), a lightweight and efficient deep learning framework designed to enhance segmentation accuracy while maintaining computational efficiency. Specifically, we introduce the Multi-Scale Spatial-Channel Attention (MSCA) module to recalibrate spatial and channel dependencies, improving local–global feature fusion while reducing redundant computations. Additionally, the Edge-Enhanced Bottleneck Attention (EEBA) module integrates Sobel-based edge detection to refine boundary delineation, ensuring sharper object separation in dense vegetation environments. Extensive evaluations on publicly available datasets demonstrate the effectiveness of MSEA-Net, achieving a mean Intersection over Union (IoU) of 87.42% on the Motion-Blurred UAV Images of Sorghum Fields dataset and 71.35% on the CoFly-WeedDB dataset, outperforming benchmark models. MSEA-Net also maintains a compact architecture with only 6.74 M parameters and a model size of 25.74 MB, making it suitable for UAV-based real-time weed segmentation. These results highlight the potential of MSEA-Net for improving automated weed detection in precision agriculture while ensuring computational efficiency for edge deployment.https://www.mdpi.com/2624-7402/7/4/103deep learningprecision agricultureweed segmentationunmanned aerial vehicle (UAV)weed management
spellingShingle Akram Syed
Baifan Chen
Adeel Ahmed Abbasi
Sharjeel Abid Butt
Xiaoqing Fang
MSEA-Net: Multi-Scale and Edge-Aware Network for Weed Segmentation
AgriEngineering
deep learning
precision agriculture
weed segmentation
unmanned aerial vehicle (UAV)
weed management
title MSEA-Net: Multi-Scale and Edge-Aware Network for Weed Segmentation
title_full MSEA-Net: Multi-Scale and Edge-Aware Network for Weed Segmentation
title_fullStr MSEA-Net: Multi-Scale and Edge-Aware Network for Weed Segmentation
title_full_unstemmed MSEA-Net: Multi-Scale and Edge-Aware Network for Weed Segmentation
title_short MSEA-Net: Multi-Scale and Edge-Aware Network for Weed Segmentation
title_sort msea net multi scale and edge aware network for weed segmentation
topic deep learning
precision agriculture
weed segmentation
unmanned aerial vehicle (UAV)
weed management
url https://www.mdpi.com/2624-7402/7/4/103
work_keys_str_mv AT akramsyed mseanetmultiscaleandedgeawarenetworkforweedsegmentation
AT baifanchen mseanetmultiscaleandedgeawarenetworkforweedsegmentation
AT adeelahmedabbasi mseanetmultiscaleandedgeawarenetworkforweedsegmentation
AT sharjeelabidbutt mseanetmultiscaleandedgeawarenetworkforweedsegmentation
AT xiaoqingfang mseanetmultiscaleandedgeawarenetworkforweedsegmentation