Multi-class segmentation of navel orange surface defects based on improved DeepLabv3+
To address the problems of current mainstream semantic segmentation network such as rough edge segmentation of navel oranges defects, poor accuracy of small target defect segmentation and insufficient deep-level semantic extraction of defects, feature information will be lost, a multi-class segment...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
PAGEPress Publications
2024-02-01
|
| Series: | Journal of Agricultural Engineering |
| Subjects: | |
| Online Access: | https://www.agroengineering.org/jae/article/view/1564 |
| _version_ | 1827346416777822208 |
|---|---|
| author | Yun Zhu Shuwen Liu Xiaojun Wu Lianfeng Gao Youyun Xu |
| author_facet | Yun Zhu Shuwen Liu Xiaojun Wu Lianfeng Gao Youyun Xu |
| author_sort | Yun Zhu |
| collection | DOAJ |
| description |
To address the problems of current mainstream semantic segmentation network such as rough edge segmentation of navel oranges defects, poor accuracy of small target defect segmentation and insufficient deep-level semantic extraction of defects, feature information will be lost, a multi-class segmentation model based on improved DeepLabv3+ is proposed to detect the surface defects of navel oranges. The Coordinate Attention Mechanism is embedded into the DeepLabv3+ network for better semantic segmentation performance, while the dilated convolution of Atrous Spatial Pyramid Pooling structure is replaced with deformable empty convolution to improve the fitting ability of the network to target shape changes and irregular defects. In addition, a BiFPN-based feature fusion branch is introduced at the DeepLabv3+ encoder side to realize multi-scale feature fusion and enrich feature space and semantic information. The experimental results show that the average intersection ratio and average pixel intersection ratio accuracies of the improved DeepLabv3+ model on the navel orange surface defect dataset are 77.32% and 86.38%, which are 3.81% and 5.29% higher than the original DeepLabv3+ network, respectively, improving the extraction capability of navel orange defect features and having better segmentation performance.
|
| first_indexed | 2024-03-07T23:30:49Z |
| format | Article |
| id | doaj.art-9f6cd1e434c4498d959b9528d3718276 |
| institution | Directory Open Access Journal |
| issn | 1974-7071 2239-6268 |
| language | English |
| last_indexed | 2024-03-07T23:30:49Z |
| publishDate | 2024-02-01 |
| publisher | PAGEPress Publications |
| record_format | Article |
| series | Journal of Agricultural Engineering |
| spelling | doaj.art-9f6cd1e434c4498d959b9528d37182762024-02-20T16:23:21ZengPAGEPress PublicationsJournal of Agricultural Engineering1974-70712239-62682024-02-0110.4081/jae.2024.1564Multi-class segmentation of navel orange surface defects based on improved DeepLabv3+Yun Zhu0Shuwen Liu1Xiaojun Wu2Lianfeng Gao3Youyun Xu4School of Physics and Electronic Information, Gannan Normal University, Ganzhou, JiangxiSchool of Physics and Electronic Information, Gannan Normal University, Ganzhou, JiangxiSchool of Physics and Electronic Information, Gannan Normal University, Ganzhou, JiangxiSchool of Physics and Electronic Information, Gannan Normal University, Ganzhou, JiangxiSchool of Communication and Information Engineering, University of Posts and Telecommunications, Nanjing, Jiangsu To address the problems of current mainstream semantic segmentation network such as rough edge segmentation of navel oranges defects, poor accuracy of small target defect segmentation and insufficient deep-level semantic extraction of defects, feature information will be lost, a multi-class segmentation model based on improved DeepLabv3+ is proposed to detect the surface defects of navel oranges. The Coordinate Attention Mechanism is embedded into the DeepLabv3+ network for better semantic segmentation performance, while the dilated convolution of Atrous Spatial Pyramid Pooling structure is replaced with deformable empty convolution to improve the fitting ability of the network to target shape changes and irregular defects. In addition, a BiFPN-based feature fusion branch is introduced at the DeepLabv3+ encoder side to realize multi-scale feature fusion and enrich feature space and semantic information. The experimental results show that the average intersection ratio and average pixel intersection ratio accuracies of the improved DeepLabv3+ model on the navel orange surface defect dataset are 77.32% and 86.38%, which are 3.81% and 5.29% higher than the original DeepLabv3+ network, respectively, improving the extraction capability of navel orange defect features and having better segmentation performance. https://www.agroengineering.org/jae/article/view/1564BiFPNcoordinate attention mechanismdeformable convolutionnavel orange surface defectssemantic segmentation |
| spellingShingle | Yun Zhu Shuwen Liu Xiaojun Wu Lianfeng Gao Youyun Xu Multi-class segmentation of navel orange surface defects based on improved DeepLabv3+ Journal of Agricultural Engineering BiFPN coordinate attention mechanism deformable convolution navel orange surface defects semantic segmentation |
| title | Multi-class segmentation of navel orange surface defects based on improved DeepLabv3+ |
| title_full | Multi-class segmentation of navel orange surface defects based on improved DeepLabv3+ |
| title_fullStr | Multi-class segmentation of navel orange surface defects based on improved DeepLabv3+ |
| title_full_unstemmed | Multi-class segmentation of navel orange surface defects based on improved DeepLabv3+ |
| title_short | Multi-class segmentation of navel orange surface defects based on improved DeepLabv3+ |
| title_sort | multi class segmentation of navel orange surface defects based on improved deeplabv3 |
| topic | BiFPN coordinate attention mechanism deformable convolution navel orange surface defects semantic segmentation |
| url | https://www.agroengineering.org/jae/article/view/1564 |
| work_keys_str_mv | AT yunzhu multiclasssegmentationofnavelorangesurfacedefectsbasedonimproveddeeplabv3 AT shuwenliu multiclasssegmentationofnavelorangesurfacedefectsbasedonimproveddeeplabv3 AT xiaojunwu multiclasssegmentationofnavelorangesurfacedefectsbasedonimproveddeeplabv3 AT lianfenggao multiclasssegmentationofnavelorangesurfacedefectsbasedonimproveddeeplabv3 AT youyunxu multiclasssegmentationofnavelorangesurfacedefectsbasedonimproveddeeplabv3 |