基于改进 YOLOv8n的海棠叶片病害检测方法

Detection Algorithm for Crabapple Leaf Diseases Based on Improved YOLOv8n

摘要

本文针对观赏植物病虫害识别的难点问题，以北方常见的观赏植物海棠树为例，提出了一种基于改进的YOLOv8n（MCSW-YOLOv8）海棠叶片常见病害的检测方法。该方法以自然环境下的海棠叶片为研究对象，对采集的样本进行增强处理。对模型进行以下优化：将原模型的主干网络替换为MobileNetV4，采用Wise-IoU（WIoU）V3作为边界框回归的损失函数，将SPP与ELAN结合起来，提高了模型对不同尺度物体的识别能力。最后添加CA注意力机制分解水平或垂直方向的池化，保留位置信息，提升目标检测中的边界框定位精准度。实验结果表明：本文提出的MCSW-YOLOv8目标检测算法在数据集上的查准率提高了7.32%，mAP@0.5提高了7.03%，mAP@0.5：0.95提高了3.53%，取得了较为理想的检测结果。总的来说，MCSW-YOLOv8模型适用于常见海棠叶病害的小目标检测。

Abstract

Aiming at the difficulties in the identification of ornamental plant pests and diseases, taking crabapple trees as example, which are common ornamental species in northern China, this paper proposes a detection method for common leaf diseases based on the improved YOLOv8n (MCSW-YOLOv8). This method targets crabapple leaves in natural environments, and enhances the collected samples. The following optimizations are applied to the model: Replace the backbone network of the original model with MobileNetV4, adopt Wise-IoU (WIoU) V3 as the loss function for bounding box regression, and integrate SPP and ELAN to enhance the model's capability to recognize objects of varying scales. Finally, incorporate CA attention mechanism to decompose horizontal and vertical pooling, retain position information, and improve the accuracy of boundary frame positioning in target detection. The results show that the MCSW-YOLOv8 object detection algorithm proposed in this paper improves the precision by 7.32%, mAP@0.5 by 7.03%, and mAP@0.5:0.95 by 3.53% on the dataset, achieving satisfactory detection performance.

关键词

深度学习 / YOLOv8n / 目标检测 / 海棠叶片病害

Key words

Deep learning / YOLOv8n / object detection / crabapple leaf diseases

参考文献

[1] 张海红. 观赏海棠病虫害发生种类及防治技术[J]. 现代农村科技, 2024(03): 55-56.

[2] 翟肇裕, 曹益飞, 徐焕良, 等. 农作物病虫害识别关键技术研究综述[J]. 农业机械学报, 2021, 52(07): 1-18.

[3] Chuanlei Z, Shanwen Z, Jucheng Y ,et al . Apple leaf disease identification using genetic algorithm and correlation based feature selection method[J]. International Journal of Agricultural and Biological Engineering, 2017, 10: 74-83.

[4] Kitsiranuwat S, Kawichai T, Khanarsa P . Identification and classification of diseases based on object detection and majority voting of bounding boxes[J]. Journal of Advances in Information Technology, 2023, 14(6): 1301-1311.

[5] Yu H J, Son C H, Lee D H . Apple leaf disease identification through region-of-interest-aware deep convolutional neural network[J]. Journal of Imaging Science and Technology, 2020, 64(2): 10.

[6] Zhou C, Xing J . Improved deep residual network for apple leaf disease identification[J]. Journal of Information Processing Systems, 2021, 17(6): 12.

[7] Ni J . Smart agriculture: An intelligent approach for apple leaf disease identification based on convolutional neural network[J]. Journal of Phytopathology, 2024(4): 172.

[8] Gao Y, Cao Z, Cai W, et al. Apple leaf disease identification in complex background based on BAM-Net[J]. Agronomy, 2023.

[9] Li L, Zhang S, Wang B . Apple leaf disease identification with a small and imbalanced dataset based on lightweight convolutional networks[J]. Sensors (Basel, Switzerland), 2021, 22(1): 2-13.

[10] 蒲攀, 刘勇, 张越, 等 . 融合动态词典特征和CBAM的苹果病虫害命名实体识别方法[J]. 农业机械学报, 2024, 55(12): 333-343.

[11] Lv M, Su W H . YOLOV5-CBAM-C3TR:An optimized model based on transformer module and attention mechanism for apple leaf disease detection[J]. Frontiers in Plant Science, 2024, 14(1): 1323301.

[12] 孙丰刚, 王云露, 兰鹏, 等. 基于改进YOLOv5s和迁移学习的苹果果实病害识别方法[J]. 农业工程学报, 2022, 38(11): 171-179.

[13] Feng L, Liu Y, Yang H, et al. Detctin of apple leaf diseases target based on improved YOLOv7[J]. INMATEH-Agricultural Engineering, 2024, 72(1): 280.

[14] 张书贵, 陈书理, 赵展. 改进YOLOv8的农作物叶片病虫害识别算法[J]. 中国农机化学报, 2024, 45(07): 255-260.

[15] Zeng W, Pang J, Ni K, et al. Apple leaf disease detection based on lightweight YOLOv8-GSSW[J]. Applied Engineering in Agriculture, 2024, 40(5): 589-598.

[16] Terven J, Cordova-Esparza D M, Romero-Gonzalez J A . A comprehensive review of YOLO architectures in computer vision: From YOLOv1 to YOLOv8 and YOLO-NAS[J]. Machine Learning And Knowledge Extraction, 2023, 5(4): 1680-1716.

[17] Qin D, Leichner C, Delakis M , et al. MobileNetV4——universal models for the mobile ecosystem[C]. European conference on computer vision. Cham: Springer Nature Switzerland, 2024: 78-96.

[18] Lu J, Zhu M, Ma X, et al. Steel strip surface defect detectionmethod based on improved YOLOv5s [J]. Biomimetics, 2024, 9(1) : 28.

[19] Wang C Y, Bochkovskiy A, Liao H Y M . YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time objectdetectors[C]. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023: 7464-7475.

[20] Hou Q, Zhou D, Feng J . Coordinate attention for efficient mobile network design[C]. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Electr Network: IEEE, 2021: 13713-13722.

[21] Tong Z J, Chen Y H, Xu Z W, et al. Wise-IoU: Bounding box regression loss with dynamic focusing mechanism[EB/OL]. 2023-01-24. https://arxiv.org/abs/2301.10051.

[22] Li C Y, Li L L, Jiang H L, et al. YOLOv6:A single-stage object detection framework for industrial applications[J]. ArXiv, 2022(9): 2209.02976.

[23] Wang C Y, Yeh I H, Liao H Y . YOLOv9: Learning what you want to learn using programmable gradientinformation[EB/OL]. 2024-02-21. https: //arxiv.org/abs/2402.13616v2.

[24] Carion N, Massan F, Synnaeve G, et al. End-to-endobject detection with transformers[C]. Proceedings of the16th European Conference on Computer Vision-ECCV 2020,Glasgow, UK, 2020: 213-229.

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

[26] Liu W, Anguelov D, Erhan D, et al. Ssd: Single shot multibox detector[C]. Computer Vision-ECCV: 14th European Conference, Amsterdam. The Netherlands: Springer International Publishing, 2016: 21-37.