Recent Advances in Plant Disease Severity Assessment Using Convolutional Neural Networks

Abstract

Abstract In modern agricultural production, the severity of diseases is an important factor that directly affects the yield and quality of plants. In order to effectively monitor and control the whole production process of crops, it is necessary to clarify not only the types of diseases, but also the severity of diseases. In recent years, the use of Deep Learning (DL) for plant disease species identification has been widely applied. In particular, the application of Convolutional Neural Networks (CNNs) for plant disease images has made breakthrough progress. However, little research has been done on disease severity assessment. This group first traces the mainstream views of existing disease research scholars and accordingly gives various scales for visual assessment of plant disease severity grading. According to the difference of network architecture, then, this study outlines the 16 researches on CNN-based plant disease severity assessment from three aspects of classical CNN framework, improved CNN architecture and CNN-based semantic segmentation network, and the advantages and disadvantages of each research method are compared and analyzed in detail. The common methods of datasets acquisition and the performance evaluation index of CNN models are examined in depth. Finally, this study discusses the major challenges in the practical application of CNN-based plant disease severity assessment methods, and provides feasible research ideas and possible solutions to address these challenges.

Code and data availability

This is a review article on CNN-based plant disease severity assessment. It presents no original phenotyping measurements, images, code, or models of its own; the authors state only that 'All data generated or analysed during this study are included in this published article and its supplementary information files' (an

No evidence-backed public reproduction asset is currently recorded.