A Multi-Strategy Framework for Enhancing Shatian Pomelo Detection in Real-World Orchards
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866918326397042688 |
|---|---|
| author | Wang, Pan Hu, Yihao Bai, Xiaodong Yang, Jingchu Zhou, Leyi Yang, Aiping Li, Xiangxiang Ding, Meiping Yao, Jianguo |
| author_facet | Wang, Pan Hu, Yihao Bai, Xiaodong Yang, Jingchu Zhou, Leyi Yang, Aiping Li, Xiangxiang Ding, Meiping Yao, Jianguo |
| contents | Shatian pomelo detection in orchards is essential for yield estimation and lean production, but models tuned to ideal datasets often degrade in practice due to device-dependent tone shifts, illumination changes, large scale variation, and frequent occlusion. We introduce STP-AgriData, a multi-scenario dataset combining real-orchard imagery with curated web images, and apply contrast/brightness augmentations to emulate unstable lighting. To better address scale and occlusion, we propose REAS-Det, featuring Global-Selective Visibility Convolution (GSV-Conv) that expands the visible feature space under global semantic guidance while retaining efficient spatial aggregation, plus C3RFEM, MultiSEAM, and Soft-NMS for refined separation and localization. On STP-AgriData, REAS-Det achieves 86.5% precision, 77.2% recall, 84.3% mAP@0.50, and 53.6% mAP@0.50:0.95, outperforming recent detectors and improving robustness in real orchard environments. The source code is available at: https://github.com/Genk641/REAS-Det. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_09948 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | A Multi-Strategy Framework for Enhancing Shatian Pomelo Detection in Real-World Orchards Wang, Pan Hu, Yihao Bai, Xiaodong Yang, Jingchu Zhou, Leyi Yang, Aiping Li, Xiangxiang Ding, Meiping Yao, Jianguo Computer Vision and Pattern Recognition Shatian pomelo detection in orchards is essential for yield estimation and lean production, but models tuned to ideal datasets often degrade in practice due to device-dependent tone shifts, illumination changes, large scale variation, and frequent occlusion. We introduce STP-AgriData, a multi-scenario dataset combining real-orchard imagery with curated web images, and apply contrast/brightness augmentations to emulate unstable lighting. To better address scale and occlusion, we propose REAS-Det, featuring Global-Selective Visibility Convolution (GSV-Conv) that expands the visible feature space under global semantic guidance while retaining efficient spatial aggregation, plus C3RFEM, MultiSEAM, and Soft-NMS for refined separation and localization. On STP-AgriData, REAS-Det achieves 86.5% precision, 77.2% recall, 84.3% mAP@0.50, and 53.6% mAP@0.50:0.95, outperforming recent detectors and improving robustness in real orchard environments. The source code is available at: https://github.com/Genk641/REAS-Det. |
| title | A Multi-Strategy Framework for Enhancing Shatian Pomelo Detection in Real-World Orchards |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2510.09948 |