Review of Wheat Disease Classification and Severity Detection Models

Abstract

Wheat is an important cereal crop that feeds more than a third of the world's population. The yield of wheat depends on various factors. Among them, disease is an important factor affecting the yield and quality of wheat. To combat these diseases, researchers have been studying the use of advanced techniques such as deep plant disease learning and image processing methods for identification. In the current study, there are many researches for wheat disease classification, but less for wheat disease severity recognition or estimate. The existing wheat disease severity detection is basically achieved by classification. Moreover, the same disease shows different symptoms at different periods or at different degrees of infection, which increases the difficulty of disease identification. In order to fully grasp the core technology of wheat disease recognition, this paper reviews the research of deep learning technology in wheat leaf disease classification and wheat disease severity. Special attention is paid to the application of image segmentation technology in wheat disease severity recognition. This paper mainly aims to explain deep learning-based wheat diseases identification algorithm, and to discuss the benefits and drawbacks of present wheat disease detection approaches. The main conclusion is that the classification of wheat diseases and the severity of wheat diseases have made good progress, but they are still in the state of independent research. Hybrid algorithm is a new way and a new challenge to link the two tasks.

Keywords

• classification
• deep learning
• disease severity
• review
• segmentation
• wheat disease