An improved U-net and attention mechanism-based model for sugar beet and weed segmentation
IntroductionWeeds are a major factor affecting crop yield and quality. Accurate identification and localization of crops and weeds are essential for achieving automated weed management in precision agriculture, especially given the challenges in recognition accuracy and real-time processing in compl...
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2025-01-01
|
| Series: | Frontiers in Plant Science |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fpls.2024.1449514/full |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841543795541278720 |
|---|---|
| author | Yadong Li Ruinan Guo Rujia Li Rongbiao Ji Mengyao Wu Dinghao Chen Cong Han Ruilin Han Yongxiu Liu Yuwen Ruan Jianping Yang Jianping Yang |
| author_facet | Yadong Li Ruinan Guo Rujia Li Rongbiao Ji Mengyao Wu Dinghao Chen Cong Han Ruilin Han Yongxiu Liu Yuwen Ruan Jianping Yang Jianping Yang |
| author_sort | Yadong Li |
| collection | DOAJ |
| description | IntroductionWeeds are a major factor affecting crop yield and quality. Accurate identification and localization of crops and weeds are essential for achieving automated weed management in precision agriculture, especially given the challenges in recognition accuracy and real-time processing in complex field environments. To address this issue, this paper proposes an efficient crop-weed segmentation model based on an improved UNet architecture and attention mechanisms to enhance both recognition accuracy and processing speed.MethodsThe model adopts the encoder-decoder structure of UNet, utilizing MaxViT (Multi-Axis Vision Transformer) as the encoder to capture both global and local features within images. Additionally, CBAM (Convolutional Block Attention Module) is incorporated into the decoder as a multi-scale feature fusion module, adaptively adjusting feature map weights to enable the model to focus more accurately on the edges and textures of crops and weeds.Results and discussionExperimental results show that the proposed model achieved 84.28% mIoU and 88.59% mPA on the sugar beet dataset, representing improvements of 3.08% and 3.15% over the baseline UNet model, respectively, and outperforming mainstream models such as FCN, PSPNet, SegFormer, DeepLabv3+, and HRNet. Moreover, the model’s inference time is only 0.0559 seconds, reducing computational overhead while maintaining high accuracy. Its performance on a sunflower dataset further verifies the model’s generalizability and robustness. This study, therefore, provides an efficient and accurate solution for crop-weed segmentation, laying a foundation for future research on automated crop and weed identification. |
| format | Article |
| id | doaj-art-dca311b612e94a39a32153b9986124a3 |
| institution | Kabale University |
| issn | 1664-462X |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Plant Science |
| spelling | doaj-art-dca311b612e94a39a32153b9986124a32025-01-13T06:11:05ZengFrontiers Media S.A.Frontiers in Plant Science1664-462X2025-01-011510.3389/fpls.2024.14495141449514An improved U-net and attention mechanism-based model for sugar beet and weed segmentationYadong Li0Ruinan Guo1Rujia Li2Rongbiao Ji3Mengyao Wu4Dinghao Chen5Cong Han6Ruilin Han7Yongxiu Liu8Yuwen Ruan9Jianping Yang10Jianping Yang11College of Big Data, Yunnan Agricultural University, Kunming, ChinaModern Educational Technology Center, Yunnan Agricultural University, Kunming, ChinaSchool of Information Technology and Artificial Intelligence, Zhejiang University of Finance and Economics, Hangzhou, ChinaCollege of Big Data, Yunnan Agricultural University, Kunming, ChinaCollege of Big Data, Yunnan Agricultural University, Kunming, ChinaCollege of Big Data, Baoshan University, Baoshan, ChinaCollege of Economics and Management, Yunnan Agricultural University, Kunming, ChinaCollege of Economics and Management, Yunnan Agricultural University, Kunming, ChinaCollege of plant protection, Yunnan Agricultural University, Kunming, ChinaFaculty of Sciences, University of Copenhagen, Copenhagen, DenmarkCollege of Big Data, Yunnan Agricultural University, Kunming, ChinaThe Key Laboratory for Crop Production and Smart Agriculture of Yunnan Province, Yunnan Agricultural University, Kunming, ChinaIntroductionWeeds are a major factor affecting crop yield and quality. Accurate identification and localization of crops and weeds are essential for achieving automated weed management in precision agriculture, especially given the challenges in recognition accuracy and real-time processing in complex field environments. To address this issue, this paper proposes an efficient crop-weed segmentation model based on an improved UNet architecture and attention mechanisms to enhance both recognition accuracy and processing speed.MethodsThe model adopts the encoder-decoder structure of UNet, utilizing MaxViT (Multi-Axis Vision Transformer) as the encoder to capture both global and local features within images. Additionally, CBAM (Convolutional Block Attention Module) is incorporated into the decoder as a multi-scale feature fusion module, adaptively adjusting feature map weights to enable the model to focus more accurately on the edges and textures of crops and weeds.Results and discussionExperimental results show that the proposed model achieved 84.28% mIoU and 88.59% mPA on the sugar beet dataset, representing improvements of 3.08% and 3.15% over the baseline UNet model, respectively, and outperforming mainstream models such as FCN, PSPNet, SegFormer, DeepLabv3+, and HRNet. Moreover, the model’s inference time is only 0.0559 seconds, reducing computational overhead while maintaining high accuracy. Its performance on a sunflower dataset further verifies the model’s generalizability and robustness. This study, therefore, provides an efficient and accurate solution for crop-weed segmentation, laying a foundation for future research on automated crop and weed identification.https://www.frontiersin.org/articles/10.3389/fpls.2024.1449514/fullsemantic segmentationUNETdeep learningMaxViTCBAMattention mechanism |
| spellingShingle | Yadong Li Ruinan Guo Rujia Li Rongbiao Ji Mengyao Wu Dinghao Chen Cong Han Ruilin Han Yongxiu Liu Yuwen Ruan Jianping Yang Jianping Yang An improved U-net and attention mechanism-based model for sugar beet and weed segmentation Frontiers in Plant Science semantic segmentation UNET deep learning MaxViT CBAM attention mechanism |
| title | An improved U-net and attention mechanism-based model for sugar beet and weed segmentation |
| title_full | An improved U-net and attention mechanism-based model for sugar beet and weed segmentation |
| title_fullStr | An improved U-net and attention mechanism-based model for sugar beet and weed segmentation |
| title_full_unstemmed | An improved U-net and attention mechanism-based model for sugar beet and weed segmentation |
| title_short | An improved U-net and attention mechanism-based model for sugar beet and weed segmentation |
| title_sort | improved u net and attention mechanism based model for sugar beet and weed segmentation |
| topic | semantic segmentation UNET deep learning MaxViT CBAM attention mechanism |
| url | https://www.frontiersin.org/articles/10.3389/fpls.2024.1449514/full |
| work_keys_str_mv | AT yadongli animprovedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT ruinanguo animprovedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT rujiali animprovedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT rongbiaoji animprovedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT mengyaowu animprovedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT dinghaochen animprovedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT conghan animprovedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT ruilinhan animprovedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT yongxiuliu animprovedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT yuwenruan animprovedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT jianpingyang animprovedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT jianpingyang animprovedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT yadongli improvedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT ruinanguo improvedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT rujiali improvedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT rongbiaoji improvedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT mengyaowu improvedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT dinghaochen improvedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT conghan improvedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT ruilinhan improvedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT yongxiuliu improvedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT yuwenruan improvedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT jianpingyang improvedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation AT jianpingyang improvedunetandattentionmechanismbasedmodelforsugarbeetandweedsegmentation |