Improved MobileVit deep learning algorithm based on thermal images to identify the water state in cotton

Thermal imaging combined with deep learning algorithms offers an efficient and non-invasive method for monitoring crop water status, facilitating precise irrigation management over large agricultural areas. This study introduces a method for identifying the moisture state of cotton using an enhanced...

Full description

Saved in:
Bibliographic Details
Main Authors: Kaijun Jin, Jihong Zhang, Ningning Liu, Miao Li, Zhanli Ma, Zhenhua Wang, Jinzhu Zhang, Feihu Yin
Format: Article
Language:English
Published: Elsevier 2025-04-01
Series:Agricultural Water Management
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S0378377425000794
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849764423130939392
author Kaijun Jin
Jihong Zhang
Ningning Liu
Miao Li
Zhanli Ma
Zhenhua Wang
Jinzhu Zhang
Feihu Yin
author_facet Kaijun Jin
Jihong Zhang
Ningning Liu
Miao Li
Zhanli Ma
Zhenhua Wang
Jinzhu Zhang
Feihu Yin
author_sort Kaijun Jin
collection DOAJ
description Thermal imaging combined with deep learning algorithms offers an efficient and non-invasive method for monitoring crop water status, facilitating precise irrigation management over large agricultural areas. This study introduces a method for identifying the moisture state of cotton using an enhanced MobileVit deep learning algorithm. This approach incorporates the Efficient Channel Attention (ECA) mechanism into the Fusion component of the MobileVit model, optimizes the first convolution in the Fusion component by replacing it with Depthwise Separable Convolution (DsConv), and substitutes the Local representation with the MobileOne block. These enhancements aim to improve model performance while maintaining its compact size. A dataset of thermal images of cotton canopies representing three different water states was developed for this study. Ablation studies were performed to evaluate the effect of each modification. Grad-CAM was utilized to illustrate the final layer features of the proposed algorithm. Various deep learning models were also trained, tested, and validated, allowing for a comparative analysis of the proposed model against traditional deep learning models in identifying cotton moisture states. The results show that the F1-score of the proposed model reaches 0.9677, achieving a recognition speed of 50.370 ms while maintaining a size of 4.94 M, outperforming other classical deep learning models. The findings of this study provide technical support for the development of future precision irrigation systems. The relevant code and datasets will be made available on GitHub (https://github.com/kingcuzamu/identifying-cotton-water-state) upon publication.
format Article
id doaj-art-33fef0704dd14d7aadeaf37e638ea662
institution DOAJ
issn 1873-2283
language English
publishDate 2025-04-01
publisher Elsevier
record_format Article
series Agricultural Water Management
spelling doaj-art-33fef0704dd14d7aadeaf37e638ea6622025-08-20T03:05:09ZengElsevierAgricultural Water Management1873-22832025-04-0131010936510.1016/j.agwat.2025.109365Improved MobileVit deep learning algorithm based on thermal images to identify the water state in cottonKaijun Jin0Jihong Zhang1Ningning Liu2Miao Li3Zhanli Ma4Zhenhua Wang5Jinzhu Zhang6Feihu Yin7College of Water Conservancy & Architectural Engineering, Shihezi University, Shihezi, Xinjiang 832000, China; Key Laboratory of Modern Water-Saving Irrigation of Xinjiang Production & Construction Group, Shihezi University, Shihezi, Xinjiang 832000, China; Technology Innovation Center for Agricultural Water and Fertilizer Efficiency Equipment of Xinjiang Production & Construction Corps, Shihezi, Xinjiang 832000, China; Key Laboratory of Northwest Oasis Water-Saving Agriculture, Ministry of Agriculture and Rural Affairs, Shihezi, Xinjiang 832000, ChinaCollege of Water Conservancy & Architectural Engineering, Shihezi University, Shihezi, Xinjiang 832000, China; Key Laboratory of Modern Water-Saving Irrigation of Xinjiang Production & Construction Group, Shihezi University, Shihezi, Xinjiang 832000, China; Technology Innovation Center for Agricultural Water and Fertilizer Efficiency Equipment of Xinjiang Production & Construction Corps, Shihezi, Xinjiang 832000, China; Key Laboratory of Northwest Oasis Water-Saving Agriculture, Ministry of Agriculture and Rural Affairs, Shihezi, Xinjiang 832000, ChinaCollege of Water Conservancy & Architectural Engineering, Shihezi University, Shihezi, Xinjiang 832000, China; Key Laboratory of Modern Water-Saving Irrigation of Xinjiang Production & Construction Group, Shihezi University, Shihezi, Xinjiang 832000, China; Technology Innovation Center for Agricultural Water and Fertilizer Efficiency Equipment of Xinjiang Production & Construction Corps, Shihezi, Xinjiang 832000, China; Key Laboratory of Northwest Oasis Water-Saving Agriculture, Ministry of Agriculture and Rural Affairs, Shihezi, Xinjiang 832000, ChinaCollege of Water Conservancy & Architectural Engineering, Shihezi University, Shihezi, Xinjiang 832000, China; Key Laboratory of Modern Water-Saving Irrigation of Xinjiang Production & Construction Group, Shihezi University, Shihezi, Xinjiang 832000, China; Technology Innovation Center for Agricultural Water and Fertilizer Efficiency Equipment of Xinjiang Production & Construction Corps, Shihezi, Xinjiang 832000, China; Key Laboratory of Northwest Oasis Water-Saving Agriculture, Ministry of Agriculture and Rural Affairs, Shihezi, Xinjiang 832000, ChinaCollege of Water Conservancy & Architectural Engineering, Shihezi University, Shihezi, Xinjiang 832000, China; Key Laboratory of Modern Water-Saving Irrigation of Xinjiang Production & Construction Group, Shihezi University, Shihezi, Xinjiang 832000, China; Technology Innovation Center for Agricultural Water and Fertilizer Efficiency Equipment of Xinjiang Production & Construction Corps, Shihezi, Xinjiang 832000, China; Key Laboratory of Northwest Oasis Water-Saving Agriculture, Ministry of Agriculture and Rural Affairs, Shihezi, Xinjiang 832000, ChinaCollege of Water Conservancy & Architectural Engineering, Shihezi University, Shihezi, Xinjiang 832000, China; Key Laboratory of Modern Water-Saving Irrigation of Xinjiang Production & Construction Group, Shihezi University, Shihezi, Xinjiang 832000, China; Technology Innovation Center for Agricultural Water and Fertilizer Efficiency Equipment of Xinjiang Production & Construction Corps, Shihezi, Xinjiang 832000, China; Key Laboratory of Northwest Oasis Water-Saving Agriculture, Ministry of Agriculture and Rural Affairs, Shihezi, Xinjiang 832000, China; Corresponding authors at: College of Water Conservancy & Architectural Engineering, Shihezi University, Shihezi 832000, China.College of Water Conservancy & Architectural Engineering, Shihezi University, Shihezi, Xinjiang 832000, China; Key Laboratory of Modern Water-Saving Irrigation of Xinjiang Production & Construction Group, Shihezi University, Shihezi, Xinjiang 832000, China; Technology Innovation Center for Agricultural Water and Fertilizer Efficiency Equipment of Xinjiang Production & Construction Corps, Shihezi, Xinjiang 832000, China; Key Laboratory of Northwest Oasis Water-Saving Agriculture, Ministry of Agriculture and Rural Affairs, Shihezi, Xinjiang 832000, China; Corresponding authors at: College of Water Conservancy & Architectural Engineering, Shihezi University, Shihezi 832000, China.College of Water Conservancy & Architectural Engineering, Shihezi University, Shihezi, Xinjiang 832000, China; Xinjiang Academy of Agricultural and Reclamation Science, Shihezi 832000, ChinaThermal imaging combined with deep learning algorithms offers an efficient and non-invasive method for monitoring crop water status, facilitating precise irrigation management over large agricultural areas. This study introduces a method for identifying the moisture state of cotton using an enhanced MobileVit deep learning algorithm. This approach incorporates the Efficient Channel Attention (ECA) mechanism into the Fusion component of the MobileVit model, optimizes the first convolution in the Fusion component by replacing it with Depthwise Separable Convolution (DsConv), and substitutes the Local representation with the MobileOne block. These enhancements aim to improve model performance while maintaining its compact size. A dataset of thermal images of cotton canopies representing three different water states was developed for this study. Ablation studies were performed to evaluate the effect of each modification. Grad-CAM was utilized to illustrate the final layer features of the proposed algorithm. Various deep learning models were also trained, tested, and validated, allowing for a comparative analysis of the proposed model against traditional deep learning models in identifying cotton moisture states. The results show that the F1-score of the proposed model reaches 0.9677, achieving a recognition speed of 50.370 ms while maintaining a size of 4.94 M, outperforming other classical deep learning models. The findings of this study provide technical support for the development of future precision irrigation systems. The relevant code and datasets will be made available on GitHub (https://github.com/kingcuzamu/identifying-cotton-water-state) upon publication.http://www.sciencedirect.com/science/article/pii/S0378377425000794CottonWater state identificationComputer visionImproved MobileVit algorithmDeep learning
spellingShingle Kaijun Jin
Jihong Zhang
Ningning Liu
Miao Li
Zhanli Ma
Zhenhua Wang
Jinzhu Zhang
Feihu Yin
Improved MobileVit deep learning algorithm based on thermal images to identify the water state in cotton
Agricultural Water Management
Cotton
Water state identification
Computer vision
Improved MobileVit algorithm
Deep learning
title Improved MobileVit deep learning algorithm based on thermal images to identify the water state in cotton
title_full Improved MobileVit deep learning algorithm based on thermal images to identify the water state in cotton
title_fullStr Improved MobileVit deep learning algorithm based on thermal images to identify the water state in cotton
title_full_unstemmed Improved MobileVit deep learning algorithm based on thermal images to identify the water state in cotton
title_short Improved MobileVit deep learning algorithm based on thermal images to identify the water state in cotton
title_sort improved mobilevit deep learning algorithm based on thermal images to identify the water state in cotton
topic Cotton
Water state identification
Computer vision
Improved MobileVit algorithm
Deep learning
url http://www.sciencedirect.com/science/article/pii/S0378377425000794
work_keys_str_mv AT kaijunjin improvedmobilevitdeeplearningalgorithmbasedonthermalimagestoidentifythewaterstateincotton
AT jihongzhang improvedmobilevitdeeplearningalgorithmbasedonthermalimagestoidentifythewaterstateincotton
AT ningningliu improvedmobilevitdeeplearningalgorithmbasedonthermalimagestoidentifythewaterstateincotton
AT miaoli improvedmobilevitdeeplearningalgorithmbasedonthermalimagestoidentifythewaterstateincotton
AT zhanlima improvedmobilevitdeeplearningalgorithmbasedonthermalimagestoidentifythewaterstateincotton
AT zhenhuawang improvedmobilevitdeeplearningalgorithmbasedonthermalimagestoidentifythewaterstateincotton
AT jinzhuzhang improvedmobilevitdeeplearningalgorithmbasedonthermalimagestoidentifythewaterstateincotton
AT feihuyin improvedmobilevitdeeplearningalgorithmbasedonthermalimagestoidentifythewaterstateincotton