Research on Variety Detection of Highland Barley Seeds Based on Improved YOLOv8

As an important characteristic crop in the Qinghai-Tibet Plateau, highland barley variety detection technology plays a key role in breeding improvement, food processing, and germplasm resource management. In recent years, deep learning-based machine vision methods have provided new ideas for seed detection. However, their application to highland barley, a special crop, still faces challenges. Most existing models are designed for plain crops and struggle to adapt to the phenotypic characteristics of highland barley seeds in high-altitude environments. Additionally, there is a lack of systematic phenotypic datasets to support model training. In view of this, this paper proposes a variety detection model for highland barley grains. By introducing an attention mechanism module and conducting multi-scale feature optimization on the head network of YOLOv8, the small target detection performance is enhanced. Furthermore, a loss calculation method more suitable for the characteristics of small target detection is adopted to further improve the overall detection performance of the model. Experimental results demonstrate that the effectiveness of the improved algorithm in this paper is significant. Compared with the original YOLOv8 model, the mean average precision (mAP) is increased by 6.1 percentage points, reaching 95.1%.

Yanbo WANG, Ruokui CHANG, Yongcheng JIANG, Xin WANG, Haifeng WANG, Shide LI, and Hua LIU. Research on Variety Detection of Highland Barley Seeds Based on Improved YOLOv8. Journal of Agricultural Science and Technology A 16 (2026) 68-78, doi: 10.17265/2161-6256/2026.02.002

[1]       Xu, P., and Zhang, Z. B. 2023. “Industrialization of functional nutritious color wheat.” China Rural Science & Technology 332 (01).

[2]       Xia, X. Y., Li, F., and Xie, K. et al. 2024. “Research progress on non-destructive detection technology of crop seed vigor.” Transactions of Chinese Society of Agricultural Engineering 43 (10): 64-73.

[3]       Ding, Z. Y., Yue, X. J., and Zeng, F. G. et al. 2023. “Spectral detection of maize seed vigor based on machine learning and deep learning.” Journal of Huazhong Agricultural University 42 (3): 230-240.

[4]       Bai, W. W., Zhao, X. N., and Luo, B. et al. 2023. “Research on wheat seed germination detection method based on YOLOv5.” Acta Agriculturae Zhejiangensis 35 (2): 445-454.

[5]       Mukasa, C., Wakholi, M. A., and Faqeerzada et al. 2022. “Nondestructive discrimination of seedless from seeded watermelon seeds by using multivariate and deep learning image analysis.” Computers and Electronics in Agriculture, article id. 106799.

[6]       Wang, X., Dong, Q., and Yang, G. 2023. “YOLOv5 improved by optimized CBAM for crop pest identification.” Comput. Syst. Appl. 32 (7): 261-268.

[7]       Wang, J., Li, Q., and Fang, Z. et al. 2023. “YOLOv6-ESG: A lightweight seafood detection method.” Journal of Marine Science and Engineering 11 (8): 1623.

[8]       Liu, Q. H., Yang, X. Y., and Jie, H. et al. 2023. “Rice grain detection based on YOLO v7 fusing of GhostNetV2.” Transactions of the Chinese Society for Agricultural Machinery 54 (12): 253-260, 299.

[9]       Zhu, D., Wen, R., and Xiong, J. 2023. “Lightweight corn silk detection network incorporating with coordinate attention mechanism.” Trans. CSAE 39: 145-153.

[10]    Dong, W., and Liang, X. Y. 2024. “An algorithmic model for recognizing healthy wheat seeds based on YOLOv8.” Third International Conference on Electronic Information Engineering, Big Data, and Computer Technology (EIBDCT 2024), Vol. 13181. SPIE, 2024.

[11]    Wang, J. P., Meng, H., and Zhen, Q. G. et al. 2024. “Camellia oleifera fruit static and dynamic detection counting based on improved COF-YOLO v8n.” Transactions of the Chinese Society for Agricultural Machinery 55 (4): 193-203.

[12]    Chen, C. Q., Fan, Y. C., and Wang, L. 2022. “Logo detection based on improved Mosaic data augmentation and feature fusion.” Computer Measurement & Control 30 (10): 188-194, 201.

[13]    Yang, L. X. et al. 2021. “Simam: A simple, parameter-free attention module for convolutional neural networks.” International Conference on Machine Learning.

[14]    Hu, J., Li, S., and Gang, S. 2018. “Squeeze-and-excitation networks.” Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.

[15]    Woo, S. et al. 2018. “CBAM: Convolutional block attention module.” Proceedings of the European Conference on Computer Vision (ECCV).

[16]    Dai, X., Chen, Y., and Xiao, B. et al. 2021. “Dynamic head: Unifying object detection heads with attentions.” Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 7373-7382.

[17]    Yan, X., and Liu, F. 2023. “Dyhead-yolov5 based on improved object detection heads with attentions.” International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery (ICNC-FSKD). IEEE, pp. 1-6.

[18]    Jiang, Z., Han, Y., and Cheng, Y. et al. 2025. “An improved YOLOv8-dyhead-WiseIoU model for positioning and counting detection of grouting sleeves in a prefabricated wall.” Journal of Construction Engineering and Management 151 (4): 04025016.

[19]    Tong, Z., Chen, Y., and Xu, Z. et al. 2023. “Wise-IoU: bounding box regression loss with dynamic focusing mechanism.” arXiv preprint arXiv: 2301.10051.

[20]    Hu, D., Yu, M., and Wu, X. et al. 2024. “DGW-YOLOv8: A small insulator target detection algorithm based on deformable attention backbone and WIoU loss function.” IET image processing 18 (4): 1096-1108.

[21]    Ren, S., He, K., and Girshick, R. et al. 2017. “Faster R-CNN: Towards real-time object detection with region proposal networks.” IEEE Transactions on Pattern Analysis and Machine Intelligence 39 (6): 1137-1149.

[22]    Zhao, Y. Q., Rao, Y., and Dong, S. P. et al. 2020. “A review of deep learning object detection methods.” Journal of Image and Graphics 25 (4): 629-654.

[23]    Xu, D. G., Wang, L., and Li, F. 2021. “A review of typical deep learning-based object detection algorithms.” Computer Engineering and Applications 57 (8): 10-25.