Lightweight Fruit-Detection Algorithm for Edge Computing Applications
In recent years, deep-learning-based fruit-detection technology has exhibited excellent performance in modern horticulture research. However, deploying deep learning algorithms in real-time field applications is still challenging, owing to the relatively low image processing capability of edge devic...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2021-10-01
|
| Series: | Frontiers in Plant Science |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fpls.2021.740936/full |
| _version_ | 1818989373163569152 |
|---|---|
| author | Wenli Zhang Yuxin Liu Kaizhen Chen Huibin Li Huibin Li Yulin Duan Yulin Duan Wenbin Wu Wenbin Wu Yun Shi Yun Shi Wei Guo |
| author_facet | Wenli Zhang Yuxin Liu Kaizhen Chen Huibin Li Huibin Li Yulin Duan Yulin Duan Wenbin Wu Wenbin Wu Yun Shi Yun Shi Wei Guo |
| author_sort | Wenli Zhang |
| collection | DOAJ |
| description | In recent years, deep-learning-based fruit-detection technology has exhibited excellent performance in modern horticulture research. However, deploying deep learning algorithms in real-time field applications is still challenging, owing to the relatively low image processing capability of edge devices. Such limitations are becoming a new bottleneck and hindering the utilization of AI algorithms in modern horticulture. In this paper, we propose a lightweight fruit-detection algorithm, specifically designed for edge devices. The algorithm is based on Light-CSPNet as the backbone network, an improved feature-extraction module, a down-sampling method, and a feature-fusion module, and it ensures real-time detection on edge devices while maintaining the fruit-detection accuracy. The proposed algorithm was tested on three edge devices: NVIDIA Jetson Xavier NX, NVIDIA Jetson TX2, and NVIDIA Jetson NANO. The experimental results show that the average detection precision of the proposed algorithm for orange, tomato, and apple datasets are 0.93, 0.847, and 0.850, respectively. Deploying the algorithm, the detection speed of NVIDIA Jetson Xavier NX reaches 21.3, 24.8, and 22.2 FPS, while that of NVIDIA Jetson TX2 reaches 13.9, 14.1, and 14.5 FPS and that of NVIDIA Jetson NANO reaches 6.3, 5.0, and 8.5 FPS for the three datasets. Additionally, the proposed algorithm provides a component add/remove function to flexibly adjust the model structure, considering the trade-off between the detection accuracy and speed in practical usage. |
| first_indexed | 2024-12-20T19:37:26Z |
| format | Article |
| id | doaj.art-8a0628612a8e49a5919662efbb66e5ad |
| institution | Directory Open Access Journal |
| issn | 1664-462X |
| language | English |
| last_indexed | 2024-12-20T19:37:26Z |
| publishDate | 2021-10-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Plant Science |
| spelling | doaj.art-8a0628612a8e49a5919662efbb66e5ad2022-12-21T19:28:37ZengFrontiers Media S.A.Frontiers in Plant Science1664-462X2021-10-011210.3389/fpls.2021.740936740936Lightweight Fruit-Detection Algorithm for Edge Computing ApplicationsWenli Zhang0Yuxin Liu1Kaizhen Chen2Huibin Li3Huibin Li4Yulin Duan5Yulin Duan6Wenbin Wu7Wenbin Wu8Yun Shi9Yun Shi10Wei Guo11Department of Information, Beijing University of Technology, Beijing, ChinaDepartment of Information, Beijing University of Technology, Beijing, ChinaDepartment of Information, Beijing University of Technology, Beijing, ChinaInstitute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing, ChinaKey Laboratory of Agricultural Remote Sensing, Ministry of Agriculture, Beijing, ChinaInstitute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing, ChinaKey Laboratory of Agricultural Remote Sensing, Ministry of Agriculture, Beijing, ChinaInstitute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing, ChinaKey Laboratory of Agricultural Remote Sensing, Ministry of Agriculture, Beijing, ChinaInstitute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing, ChinaKey Laboratory of Agricultural Remote Sensing, Ministry of Agriculture, Beijing, ChinaInternational Field Phenomics Research Laboratory, Institute for Sustainable Agro-Ecosystem Services, The University of Tokyo, Tokyo, JapanIn recent years, deep-learning-based fruit-detection technology has exhibited excellent performance in modern horticulture research. However, deploying deep learning algorithms in real-time field applications is still challenging, owing to the relatively low image processing capability of edge devices. Such limitations are becoming a new bottleneck and hindering the utilization of AI algorithms in modern horticulture. In this paper, we propose a lightweight fruit-detection algorithm, specifically designed for edge devices. The algorithm is based on Light-CSPNet as the backbone network, an improved feature-extraction module, a down-sampling method, and a feature-fusion module, and it ensures real-time detection on edge devices while maintaining the fruit-detection accuracy. The proposed algorithm was tested on three edge devices: NVIDIA Jetson Xavier NX, NVIDIA Jetson TX2, and NVIDIA Jetson NANO. The experimental results show that the average detection precision of the proposed algorithm for orange, tomato, and apple datasets are 0.93, 0.847, and 0.850, respectively. Deploying the algorithm, the detection speed of NVIDIA Jetson Xavier NX reaches 21.3, 24.8, and 22.2 FPS, while that of NVIDIA Jetson TX2 reaches 13.9, 14.1, and 14.5 FPS and that of NVIDIA Jetson NANO reaches 6.3, 5.0, and 8.5 FPS for the three datasets. Additionally, the proposed algorithm provides a component add/remove function to flexibly adjust the model structure, considering the trade-off between the detection accuracy and speed in practical usage.https://www.frontiersin.org/articles/10.3389/fpls.2021.740936/fullmodern horticulturedeep learningfruit detectionlightweightedge devices |
| spellingShingle | Wenli Zhang Yuxin Liu Kaizhen Chen Huibin Li Huibin Li Yulin Duan Yulin Duan Wenbin Wu Wenbin Wu Yun Shi Yun Shi Wei Guo Lightweight Fruit-Detection Algorithm for Edge Computing Applications Frontiers in Plant Science modern horticulture deep learning fruit detection lightweight edge devices |
| title | Lightweight Fruit-Detection Algorithm for Edge Computing Applications |
| title_full | Lightweight Fruit-Detection Algorithm for Edge Computing Applications |
| title_fullStr | Lightweight Fruit-Detection Algorithm for Edge Computing Applications |
| title_full_unstemmed | Lightweight Fruit-Detection Algorithm for Edge Computing Applications |
| title_short | Lightweight Fruit-Detection Algorithm for Edge Computing Applications |
| title_sort | lightweight fruit detection algorithm for edge computing applications |
| topic | modern horticulture deep learning fruit detection lightweight edge devices |
| url | https://www.frontiersin.org/articles/10.3389/fpls.2021.740936/full |
| work_keys_str_mv | AT wenlizhang lightweightfruitdetectionalgorithmforedgecomputingapplications AT yuxinliu lightweightfruitdetectionalgorithmforedgecomputingapplications AT kaizhenchen lightweightfruitdetectionalgorithmforedgecomputingapplications AT huibinli lightweightfruitdetectionalgorithmforedgecomputingapplications AT huibinli lightweightfruitdetectionalgorithmforedgecomputingapplications AT yulinduan lightweightfruitdetectionalgorithmforedgecomputingapplications AT yulinduan lightweightfruitdetectionalgorithmforedgecomputingapplications AT wenbinwu lightweightfruitdetectionalgorithmforedgecomputingapplications AT wenbinwu lightweightfruitdetectionalgorithmforedgecomputingapplications AT yunshi lightweightfruitdetectionalgorithmforedgecomputingapplications AT yunshi lightweightfruitdetectionalgorithmforedgecomputingapplications AT weiguo lightweightfruitdetectionalgorithmforedgecomputingapplications |