Grape cultivation is affected by a variety of diseases and pests. Addressing the issues of insufficient labeled samples in existing grape leaf disease and pest identification models and limited generalization capability of the networks, this paper proposes a grape leaf disease and pest detection method assisted by image style transfer. During the preprocessing stage, gamma correction is used to reduce the impact of lighting variations in the real environment on grape leaf image detection. In the data augmentation phase, an improved cycle generative adversarial network (CycleGAN) is employed, adding a category loss to incorporate category labels during the generator's training process. Category label information for grape diseases and pests, along with images created by the generator, are input into the discriminator. This converts samples of original healthy grape leaves from various growth stages into samples that represent affected regions, enriching the data augmentation for image samples input into the detection network and increasing the variety of the samples. In the grape leaf image detection process, a visual attention mechanism module is introduced into the Backbone structure of YOLOv8 to optimize the original feature extraction network. This module assigns and dynamically adjusts different attention weights to the diseased and pest-infested areas, thus augmenting the salience of grape leaf disease features against the complete background of the leaf. Experimental results show that the improved detection algorithm achieves a 4% increase in accuracy and a 7.1% increase in recall rate on the test set for grape leaf diseases and pests compared to the baseline algorithms, providing a foundation for the selection of pesticide usage in the development of smart agriculture.

Qi Li, Xiaohui Pang. Research on Grape Leaf Disease and Pest Detection Method Based on Style Transfer Assistance. Academic Journal of Computing & Information Science (2023), Vol. 6, Issue 12: 70-78. https://doi.org/10.25236/AJCIS.2023.061208.

[1] He, X., Li, S., & Liu, B. Grape Leaf Disease Recognition Based on Multi-Scale Residual Neural Network [J]. Computer Engineering, 2021, 47(5), 285-291, 300.

[2] Guo, W., Feng, Q., Li, X., Yang, S., & Yang, J. Grape leaf disease detection based on attention mechanisms [J]. International Journal of Agricultural and Biological Engineering, 2022, 15(5):8.

[3] Ji, M., Zhang, L., & Wu, Q. Automatic grape leaf diseases identification via United Model based on multiple convolutional neural networks [J].Information Processing in Agriculture, 2020, 7(3):418-426.

[4] Y. Liu, F. E. T. Munsayac, N. T. Bugtai and R. G. Baldovino. Image Style Transfer with Feature Extraction Algorithm using Deep Learning[C]. 2021 IEEE 13th International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment, and Management (HNICEM),2021, pp. 1-5.

[5] Zhang, Y., Tian, Y., & Hou, J.Csast: content self-supervised and style contrastive learning for arbitrary style transfer. Neural Networks: The Official Journal of the International Neural Network Society, 2023, 164: 146-155.

[6] Gatys L A, Ecker A S, Bethge M.A neural algorithm of artistic style [J]. Journal of Vision, 2015.

[7] Gatys, L. A, Ecker, A. S., & Bethge, M. Image Style Transfer Using Convolutional Neural Networks[C]. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016,2414-2423.

[8] Zhu J Y, Park T, Isola P, et al. Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks[C]. 2017 IEEE International Conference on Computer Vision (ICCV), 2017, pp. 2242-2251.

[9] Woo, S., Park, J., Lee, J., & Kweon, I. (2018). CBAM: Convolutional Block Attention Module[C]. Proceedings of the European Conference on Computer Vision (ECCV), 2018, 3-19.

[10] Wang, X., Girshick, R., Gupta, A., & He, K. Non-local neural networks[C]. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2018, 7794-7803.