Efficient and accurate tobacco leaf maturity detection: an improved YOLOv10 model with DCNv3 and efficient local attention integration DOI Creative Commons
Yi Shi, Hong Wang, Fei Wang

и другие.

Frontiers in Plant Science, Год журнала: 2025, Номер 15

Опубликована: Янв. 3, 2025

The precise determination of tobacco leaf maturity is pivotal for safeguarding the taste and quality products, augmenting financial gains growers, propelling industry's sustainable progression. This research addresses inherent subjectivity variability in conventional evaluation techniques reliant on human expertise by introducing an innovative YOLOv10-based method detection. technique facilitates a rapid non-invasive assessment maturity, significantly elevating accuracy efficiency evaluation. In our study, we have advanced YOLOv10 framework integrating DCNv3 with C2f to construct enhanced neck network, designated as C2f-DCNv3. integration designed augment model's capability feature integration, particularly concerning morphological edge characteristics leaves. Furthermore, incorporation Efficient Local Attention (ELA) mechanism at multiple stages model has substantially fidelity extraction. empirical results underscore pronounced enhancement performance across all classifications. Notably, overall precision (P) been elevated from 0.939 0.973, recall rate (R) improved 0.968 0.984, mean average 50% intersection over union (mAP50) 0.984 0.994, 95% range (mAP50-95) risen 0.962 0.973. presents industry novel detection instrument endowed substantial practical utility broad prospects application. Future endeavors will be directed towards further optimization architecture bolster its generalizability explore implementation within realm actual cultivation processing.

Язык: Английский

Efficient and accurate tobacco leaf maturity detection: an improved YOLOv10 model with DCNv3 and efficient local attention integration DOI Creative Commons
Yi Shi, Hong Wang, Fei Wang

и другие.

Frontiers in Plant Science, Год журнала: 2025, Номер 15

Опубликована: Янв. 3, 2025

The precise determination of tobacco leaf maturity is pivotal for safeguarding the taste and quality products, augmenting financial gains growers, propelling industry's sustainable progression. This research addresses inherent subjectivity variability in conventional evaluation techniques reliant on human expertise by introducing an innovative YOLOv10-based method detection. technique facilitates a rapid non-invasive assessment maturity, significantly elevating accuracy efficiency evaluation. In our study, we have advanced YOLOv10 framework integrating DCNv3 with C2f to construct enhanced neck network, designated as C2f-DCNv3. integration designed augment model's capability feature integration, particularly concerning morphological edge characteristics leaves. Furthermore, incorporation Efficient Local Attention (ELA) mechanism at multiple stages model has substantially fidelity extraction. empirical results underscore pronounced enhancement performance across all classifications. Notably, overall precision (P) been elevated from 0.939 0.973, recall rate (R) improved 0.968 0.984, mean average 50% intersection over union (mAP50) 0.984 0.994, 95% range (mAP50-95) risen 0.962 0.973. presents industry novel detection instrument endowed substantial practical utility broad prospects application. Future endeavors will be directed towards further optimization architecture bolster its generalizability explore implementation within realm actual cultivation processing.

Язык: Английский

Процитировано

0