An Improved Lightweight Network for Real-Time Detection of Apple Leaf Diseases in Natural Scenes

Abstract

Achieving rapid and accurate detection of apple leaf diseases in the natural environment is essential for the growth of apple plants and the development of the apple industry. In recent years, deep learning has been widely studied and applied to apple leaf disease detection. However, existing networks have too many parameters to be easily deployed or lack research on leaf diseases in complex backgrounds to effectively use in real agricultural environments. This study proposes a novel deep learning network, YOLOX-ASSANano, which is an improved lightweight real-time model for apple leaf disease detection based on YOLOX-Nano. We improved the YOLOX-Nano backbone using a designed asymmetric ShuffleBlock, a CSP-SA module, and blueprint-separable convolution (BSConv), which significantly enhance feature-extraction capability and boost detection performance. In addition, we construct a multi-scene apple leaf disease dataset (MSALDD) for experiments. The experimental results show that the YOLOX-ASSANano model with only 0.83 MB parameters achieves 91.08% mAP on MSALDD and 58.85% mAP on the public dataset PlantDoc with a speed of 122 FPS. This study indicates that the YOLOX-ASSANano provides a feasible solution for the real-time diagnosis of apple leaf diseases in natural scenes, and could be helpful for the detection of other plant diseases.

YOLOV5-CBAM-C3TR: an optimized model based on transformer module and attention mechanism for apple leaf disease detection

“…Traditionally, the identification of apple leaf disease mostly relied on experienced farmers to identify the disease. However, due to the similarity of diseases or the complexity of symptoms, relying on human eye detection can easily lead to misjudgment of diseases, which can not only solve the problem of diseases but also cause environmental pollution ( Liu et al., 2022 ). The combination of machine learning and image processing replaced human eye detection and provided a new direction for disease detection.…”

HSSNet: A End-to-End Network for Detecting Tiny Targets of Apple Leaf Diseases in Complex Backgrounds

“…They were not tailored to design apple leaf, disease-specific network models. In 2022, the YOLOX-ASSANano [21] lightweight model based on YOLOX-Nano was designed for real-time apple disease detection. The model showed a performance of 58.85% mAP at 122 FPS on the public dataset PlantDoc.…”

Maize seed appearance quality assessment based on improved Inception-ResNet

“…Compared with traditional machine learning techniques, convolutional neural networks are more generalizable, faster to train, and can obtain significant information directly from images, which eliminates the tedious steps of manually extracting image features used in traditional methods. In applications for agriculture, convolutional neural networks are often used in areas such as the classification of crop pests and diseases ( Wu et al, 2019 ; Peng et al, 2019 ; Tiwari et al., 2021 ; Liu et al., 2022 ; Liu et al., 2022 ), agricultural product species identification ( Ajit et al., 2020 ; Gao et al., 2020 ; Chen et al, 2021 ; Laabassi et al., 2021 ; Sj et al.,2021 ), yield estimation ( Zhang et al., 2020 ; Tan et al, 2019 ; Alexandros et al, 2023 ; Kavita et al., 2023 ), and crop quality grading ( Anikó and Miklós, 2022 ; Liu et al., 2022 ; Li et al, 2022 ; Wang Z. et al., 2022 ; Peng et al, 2023 ), in which they greatly promote the development of agricultural intelligence. Along with the arrival of the era of big data, the amount of image information increases exponentially, resulting in an increase in the amount of computation and training difficulty in the training process.…”