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

Gensheng Hu,Wenxia Bao,Tao Fan,Dong Liang,Haidong Li

doi:10.1038/s41598-022-06181-z

Abstract

The accurate detection and identification of tea leaf diseases are conducive to its precise prevention and control. Convolutional neural network (CNN) can automatically extract the features of diseased tea leaves in the images. However, tea leaf images taken in natural environments have problems, such as complex backgrounds, dense leaves, and large-scale changes. The existing CNNs have low accuracy in detecting and identifying tea leaf diseases. This study proposes an improved RetinaNet target detection and identification network, AX-RetinaNet, which is used for the automatic detection and identification of tea leaf diseases in natural scene images. AX-RetinaNet uses an improved multiscale feature fusion module of the X-module and adds a channel attention module, Attention. The feature fusion module of the X-module obtains feature maps with rich information through multiple fusions of multi-scale features. The attention module assigns a network adaptively optimized weight to each feature map channel so that the network can select more effective features and reduce the interference of redundant features. This study also uses data augmentation methods to solve the problem of insufficient samples. Experimental results show the detection and identification accuracy of AX-RetinaNet for tea leaf diseases in natural scene images is better than the existing target detection and identification networks, such as SSD, RetinaNet, YOLO-v3, YOLO-v4, Centernet, M2det, and EfficientNet. The AX-RetinaNet detection and identification results indicated the mAP value of 93.83% and the F1-score value of 0.954. Compared with the original network, the mAP value, recall value, and identification accuracy increased by nearly 4%, by 4%, and by nearly 1.5%, respectively.

Highlights

The accurate detection and identification of tea leaf diseases are conducive to its precise prevention and control
Among the images of diseased tea leaves captured in the natural environments, 700 images of four types of tea leaf diseases were selected to construct a dataset, including 175 images of tea algae leaf spot (TALS), 175 images of tea bud blight (TBB), 175 images of tea white scab (TWS), and 175 images of tea leaf blight (TLB)
The results show that the channel Attention module only improves the performance of some Convolutional neural network (CNN), such as RetinaNet, You Only Look Once (YOLO)-v4, and M2det

Summary

Introduction

The accurate detection and identification of tea leaf diseases are conducive to its precise prevention and control. Convolutional neural network (CNN) can automatically extract the features of diseased tea leaves in the images. With the development of computer technology, the use of machine learning and image processing technology to achieve automatic detection and identification of crop diseases has become an important method for the automatic diagnosis of crop diseases[2,3]. Jiang et al combined deep learning and SVM methods to identify four rice diseases, among which CNN was used to extract the features of diseased rice leaves in the images, and the SVM was applied to classify and predict the specific disease[11]. Hu et al proposed a low shot learning method that uses SVM to segment the disease spots in the images of diseased tea leaves to eliminate background interference, and used an improved conditional deep convolutional generative adversarial networks(C-DCGAN) to solve the problem of insufficient s amples[12]. The multi-task problems of detection and identification are not involved, thereby greatly limiting the application of these methods

Methods

Results

Conclusion