PDF
Full
Outline
To evaluate the feasibility of using deep learning for efficient and accurate wheat spike counting, ten major winter wheat cultivars from the Huang-Huai wheat region (Fanmai 8, Zhoumai 36, Zhongmai 895, Malan 1, Xinmai 26, Yumai 49, Jimai 22, Zhongmai 578, Zhengmai 1860, and Zhongmai 255) were selected as materials. Three planting densities (1.2 million, 2.4 million, and 3.6 million plants·hm-2) were tested, and four deep learning algorithms (YOLO v5, YOLO v6, YOLO v8, and YOLO v10) were employed to construct real-time video-based wheat spike detection and counting models. The models were validated using field-grown Zhongmai 578 populations (3 million plants·hm-2). The results showed that the initial loss functions and convergence rates varied among models, with all models improving in spike detection performance as iterations increased. In terms of training speed and inference efficiency, YOLO v6 and YOLO v10 were faster but exhibited lower detection accuracy compared to YOLO v5 and YOLO v8. Although YOLO v5 and YOLO v8 required longer processing time, YOLO v8 achieved the best performance in recall (90.90%), F1-score (93.00%), mean average precision (97.20%), and overall accuracy (88.00%). The correlation (r2) between YOLO v8’s spike counts and manual counts decreased with increasing planting density, yielding values of 0.92, 0.81, and 0.79 for the three densities, respectively. Field validation demonstrated that YOLO v8 outperformed other models in stability and precision across different grain-filling stages, with the highest r2 (0.90) for real-time video-based spike counting. The model also showed robustness against variations in planting density, cultivar, and spike growth stage, maintaining high performance in complex field environments. These findings suggest that YOLO v8 is a reliable algorithm for wheat spike counting, suitable for yield prediction, breeding, and cultivation management applications.
PDF
24
9
| 科 Family | 属数 Number of genus | 种数 Number of species | 占总种数比例 Percentage of total species (%) | 属 Genus | 种数 Number of species | 占总种数比例 Percentage of total species (%) |
|---|---|---|---|---|---|---|
| 鹅膏菌科Amanitaceae | 2 | 11 | 5.26 | 鹅膏菌属 Amanita | 10 | 4.78 |
| 小菇科 Mycenaceae | 2 | 12 | 5.74 | 丝盖伞属 Inocybe | 5 | 2.39 |
| 多孔菌科 Polyporaceae | 8 | 14 | 6.70 | 蜡蘑属 Laccaria | 5 | 2.39 |
| 红菇科 Russulaceae | 3 | 23 | 11.00 | 小皮伞属 Marasmius | 6 | 2.87 |
| 小菇属 Mycena | 11 | 5.26 | ||||
| 光柄菇属 Pluteus | 5 | 2.39 | ||||
| 红菇属 Russula | 17 | 8.13 | ||||
| 栓菌属 Trametes | 5 | 2.39 |
关闭全屏
No. 86 Xueyuan South Road, Haidian District, Beijing
010-62199257
qkjq@cast.org.cn
Copyright © 2025 China Association for Science and Technology. All rights reserved. For all open access content, the relevant licensing terms apply.
Sponsored by the Office of the Leading Group for Cybersecurity and Informatization of CAST, and supported by Science and Technology Review Publishing House
