MSCPUnet: A multi-task neural network for plot-level crop classification in complex agricultural areas

High-precision mapping of agricultural crops in complex planting areas is a prerequisite for precision agricultural management. This paper first proposes a novel multi-task neural network, Multi-task Multi-Scale Convolutional Pooling Unet (MSCPUnet), for extracting vector plots. MSCPUnet is based on...

Full description

Saved in:
Bibliographic Details
Main Authors: Kedi Fang, Shengwei Zhang, Yongting Han, Lin Yang, Meng Luo, Lu Liu, Qian Zhang, Bo Wang
Format: Article
Language:English
Published: Elsevier 2024-12-01
Series:Smart Agricultural Technology
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S277237552400265X
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850244986318094336
author Kedi Fang
Shengwei Zhang
Yongting Han
Lin Yang
Meng Luo
Lu Liu
Qian Zhang
Bo Wang
author_facet Kedi Fang
Shengwei Zhang
Yongting Han
Lin Yang
Meng Luo
Lu Liu
Qian Zhang
Bo Wang
author_sort Kedi Fang
collection DOAJ
description High-precision mapping of agricultural crops in complex planting areas is a prerequisite for precision agricultural management. This paper first proposes a novel multi-task neural network, Multi-task Multi-Scale Convolutional Pooling Unet (MSCPUnet), for extracting vector plots. MSCPUnet is based on the Unet model and enhances performance through the incorporation of attention mechanisms, multi-scale pooling layers, and a multi-task learning approach for parallel processing. A ZiYuan-1 02D (ZY1E) satellite image collected from the He-Tao irrigation district in China is selected for plot extraction experiments, where MSCPUnet is compared with three other deep learning network variants. The Inter section over Union (IoU) and Accuracy metrics for the MSCPUnet model achieve the highest values, converging at 0.837 and 0.928, respectively. Leveraging this capability, a crop classification framework is proposed, which first extracts crop attributes from Sentinel-2 (S2) time series data. The plot information obtained from the MSCPUnet model is then combined with the area dominance method to assign crop attributes to vector plots, facilitating crop structure identification at the plot scale. Results indicate that this method significantly improves classification accuracy for fragmented farmlands, with overall accuracy (OA) rising to 91 % and Kappa coefficient increasing to 0.84 compared to a random forest classifier. This integrated crop classification approach has been validated in this study for high-precision crop mapping in complex planting areas.
format Article
id doaj-art-17a7669795ab4ab08dc5df5494df28f9
institution OA Journals
issn 2772-3755
language English
publishDate 2024-12-01
publisher Elsevier
record_format Article
series Smart Agricultural Technology
spelling doaj-art-17a7669795ab4ab08dc5df5494df28f92025-08-20T01:59:35ZengElsevierSmart Agricultural Technology2772-37552024-12-01910066010.1016/j.atech.2024.100660MSCPUnet: A multi-task neural network for plot-level crop classification in complex agricultural areasKedi Fang0Shengwei Zhang1Yongting Han2Lin Yang3Meng Luo4Lu Liu5Qian Zhang6Bo Wang7College of Water Conservancy and Civil Engineering, Inner Mongolia Agricultural University, Hohhot 010018, ChinaCollege of Water Conservancy and Civil Engineering, Inner Mongolia Agricultural University, Hohhot 010018, China; Inner Mongolia Key Laboratory of Ecohydrology and High Efficient Utilization of Water Resources, Hohhot 010018, China; Autonomous Region Collaborative Innovation Center for Integrated Management of Water Resources and Water Environment in the Inner Mongolia Reaches of the Yellow River, Hohhot 010018, China; Corresponding author.College of Water Conservancy and Civil Engineering, Inner Mongolia Agricultural University, Hohhot 010018, ChinaCollege of Water Conservancy and Civil Engineering, Inner Mongolia Agricultural University, Hohhot 010018, ChinaCollege of Water Conservancy and Civil Engineering, Inner Mongolia Agricultural University, Hohhot 010018, ChinaCollege of Water Conservancy and Civil Engineering, Inner Mongolia Agricultural University, Hohhot 010018, ChinaCollege of Water Conservancy and Civil Engineering, Inner Mongolia Agricultural University, Hohhot 010018, ChinaCollege of Water Conservancy and Civil Engineering, Inner Mongolia Agricultural University, Hohhot 010018, ChinaHigh-precision mapping of agricultural crops in complex planting areas is a prerequisite for precision agricultural management. This paper first proposes a novel multi-task neural network, Multi-task Multi-Scale Convolutional Pooling Unet (MSCPUnet), for extracting vector plots. MSCPUnet is based on the Unet model and enhances performance through the incorporation of attention mechanisms, multi-scale pooling layers, and a multi-task learning approach for parallel processing. A ZiYuan-1 02D (ZY1E) satellite image collected from the He-Tao irrigation district in China is selected for plot extraction experiments, where MSCPUnet is compared with three other deep learning network variants. The Inter section over Union (IoU) and Accuracy metrics for the MSCPUnet model achieve the highest values, converging at 0.837 and 0.928, respectively. Leveraging this capability, a crop classification framework is proposed, which first extracts crop attributes from Sentinel-2 (S2) time series data. The plot information obtained from the MSCPUnet model is then combined with the area dominance method to assign crop attributes to vector plots, facilitating crop structure identification at the plot scale. Results indicate that this method significantly improves classification accuracy for fragmented farmlands, with overall accuracy (OA) rising to 91 % and Kappa coefficient increasing to 0.84 compared to a random forest classifier. This integrated crop classification approach has been validated in this study for high-precision crop mapping in complex planting areas.http://www.sciencedirect.com/science/article/pii/S277237552400265XMulti-task neural networkPlot segmentationHigh-precision crop mappingZY1E satellite
spellingShingle Kedi Fang
Shengwei Zhang
Yongting Han
Lin Yang
Meng Luo
Lu Liu
Qian Zhang
Bo Wang
MSCPUnet: A multi-task neural network for plot-level crop classification in complex agricultural areas
Smart Agricultural Technology
Multi-task neural network
Plot segmentation
High-precision crop mapping
ZY1E satellite
title MSCPUnet: A multi-task neural network for plot-level crop classification in complex agricultural areas
title_full MSCPUnet: A multi-task neural network for plot-level crop classification in complex agricultural areas
title_fullStr MSCPUnet: A multi-task neural network for plot-level crop classification in complex agricultural areas
title_full_unstemmed MSCPUnet: A multi-task neural network for plot-level crop classification in complex agricultural areas
title_short MSCPUnet: A multi-task neural network for plot-level crop classification in complex agricultural areas
title_sort mscpunet a multi task neural network for plot level crop classification in complex agricultural areas
topic Multi-task neural network
Plot segmentation
High-precision crop mapping
ZY1E satellite
url http://www.sciencedirect.com/science/article/pii/S277237552400265X
work_keys_str_mv AT kedifang mscpunetamultitaskneuralnetworkforplotlevelcropclassificationincomplexagriculturalareas
AT shengweizhang mscpunetamultitaskneuralnetworkforplotlevelcropclassificationincomplexagriculturalareas
AT yongtinghan mscpunetamultitaskneuralnetworkforplotlevelcropclassificationincomplexagriculturalareas
AT linyang mscpunetamultitaskneuralnetworkforplotlevelcropclassificationincomplexagriculturalareas
AT mengluo mscpunetamultitaskneuralnetworkforplotlevelcropclassificationincomplexagriculturalareas
AT luliu mscpunetamultitaskneuralnetworkforplotlevelcropclassificationincomplexagriculturalareas
AT qianzhang mscpunetamultitaskneuralnetworkforplotlevelcropclassificationincomplexagriculturalareas
AT bowang mscpunetamultitaskneuralnetworkforplotlevelcropclassificationincomplexagriculturalareas