Abstract:To address the issues of low accuracy and slow response speed in tea disease classification and identification, an improved YOLOv7 lightweight model was proposed in this study. The lightweight MobileNeXt was used as the backbone network to reduce computational load and enhance efficiency. Additionally, a dual-layer routing attention mechanism was introduced to enhance the model's ability to capture crucial details and textures in disease images, thereby improving accuracy. The SIoU loss function was employed to mitigate missed and erroneous judgments, resulting in improved recognition amidst complex image backgrounds.The revised model achieved precision, recall, and average precision of 93.5%, 89.9%, and 92.1%, respectively, representing increases of 4.5%, 1.9%, and 2.6% over the original model. Furthermore, the model's volum was reduced by 24.69M, the total param was reduced by 12.88M, while detection speed was increased by 24.41 frames per second. This enhanced model efficiently and accurately identifies tea disease types, offering the benefits of lower parameter count and faster detection, thereby establishing a robust foundation for tea disease monitoring and prevention efforts.

What problem does this paper attempt to address?

The paper attempts to address the issues of low accuracy and slow response speed in the classification and identification of tea plant diseases. Specifically, the authors propose an improved lightweight YOLOv7 model to enhance the accuracy and efficiency of tea plant disease classification and identification. Below is a summary of the main content of the paper: ### Research Background - **Impact of Tea Plant Diseases**: The long-term warm and humid environment leads to a variety of tea plant diseases, making prevention and control difficult. This severely affects tea yield and quality, causing significant economic losses to tea farmers and producers. - **Limitations of Existing Methods**: Traditional object detection algorithms rely on shallow features such as color and texture, making it difficult to effectively extract deep features. This results in poor robustness and generalization ability in complex scenarios. ### Research Objectives - **Improve Accuracy and Response Speed**: By improving the YOLOv7 model, the computational load is reduced, and the model's ability to recognize in complex backgrounds is enhanced. - **Reduce Parameter Count and Increase Detection Speed**: Achieve model lightweighting to make it more suitable for deployment and application on mobile devices. ### Methods 1. **Dataset**: - The dataset comes from the Yunnan Academy of Agricultural Sciences and the Hekai Tea Garden Base in Menghai County, containing 1823 images with a resolution of 4608×3456 pixels. - Data augmentation techniques (such as color enhancement, flipping, random rotation, etc.) are used to expand the dataset, resulting in a final total of 4280 images. 2. **Improved YOLOv7 Model**: - **MobileNeXt as Backbone Network**: Replaces the original CSPDarknet53, reducing model parameters and improving inference speed. - **Dual-Layer Routing Attention Mechanism**: Introduces this mechanism to enhance the model's ability to capture key details and textures in disease images, improving robustness and generalization ability. - **SIoU Loss Function**: Optimizes the loss function to make the predicted box approach the real box faster, reducing regression loss and accelerating model convergence. ### Experimental Results - **Performance Improvement**: - The improved YOLOv7 model increased precision, recall, and mean average precision by 4.5%, 1.9%, and 2.6%, respectively. - The model parameter count was reduced by 12.88M, and detection speed increased by 24.41 frames per second. - **Comparison with Other Models**: - On the same test dataset, the improved YOLOv7 model significantly outperformed other models (such as YOLOv7, YOLOv7-tiny, Faster R-CNN, and SSD) in terms of mAP metrics. ### Conclusion - **Model Advantages**: The improved YOLOv7 model demonstrates higher accuracy and faster detection speed in the classification and identification of tea plant diseases in complex backgrounds, providing a reliable foundation for tea plant disease monitoring and prevention. Through these improvements, this study provides a new solution for the efficient and accurate identification of tea plant diseases, contributing to tea plant management and sustainable development.

Classification and identification of tea diseases based on improved YOLOv7 model of MobileNeXt

Detection Model of Tea Disease Severity under Low Light Intensity Based on YOLOv8 and EnlightenGAN

YOLOv8-RMDA: Lightweight YOLOv8 Network for Early Detection of Small Target Diseases in Tea

YOLO-Tea: A Tea Disease Detection Model Improved by YOLOv5

Tea leaf disease and insect identification based on improved MobileNetV3

Tea leaf disease detection and identification based on YOLOv7 (YOLO-T)

Automated Tea Leaves Recognition: Multi-Classification Using YOLOv5 and Inception V3 Model

Detection of Small Foreign Objects in Pu-erh Sun-Dried Green Tea: an Enhanced YOLOv8 Neural Network Model Based on Deep Learning

Identification of Tea Disease Under Complex Backgrounds Based on Minimalism Neural Network Architecture and Channel Reconstruction Unit

Research on lightweight pavement disease detection model based on YOLOv7

Deep migration learning-based recognition of diseases and insect pests in Yunnan tea under complex environments

Integrated Learning-Based Pest and Disease Detection Method for Tea Leaves

Pest recognition in microstates state: an improvement of YOLOv7 based on Spatial and Channel Reconstruction Convolution for feature redundancy and vision transformer with Bi-Level Routing Attention

Detection and identification of tea leaf diseases based on AX-RetinaNet

Lightweight Network for Corn Leaf Disease Identification Based on Improved YOLO V8s

Improved YOLOv8 and SAHI Model for the Collaborative Detection of Small Targets at the Micro Scale: A Case Study of Pest Detection in Tea

Advanced deep learning algorithm for instant discriminating of tea leave stress symptoms by smartphone-based detection

UAV remote sensing detection of tea leaf blight based on DDMA-YOLO

REAL-TIME GRAPE DISEASE DETECTION MODEL BASED ON IMPROVED YOLOv8s

A Small Target Tea Leaf Disease Detection Model Combined with Transfer Learning

Plant Disease Recognition Model Based on Improved YOLOv5