Published Paper Details:

Plant diseases and pests are major factors that dictate yield and quality of plants. Identification of plant diseases and pests can be done using digital image processing. Recent years have brought a breakthrough in digital image processing by deep learning that far outstrips what can be achieved with traditional techniques. One big issue among researchers is that developing deep learning technology is going into research on plant diseases and pests identification. This review provides the definition of the problem in plant diseases and pests detection and compares traditional methods used in plant diseases and pests detection. Based on the differences in the network structure, this paper summarizes research on plant disease and pest detection using deep learning over the last few years, which include classification networks, detection networks, and segmentation networks, and summarize their advantages and disadvantages. Common datasets are introduced, and performance of existing studies is compared. Based on this, the paper will introduce possible challenges for practical application in deep learning-based plant diseases and pests detection. Additionally, possible solutions and research ideas concerning the challenges will be proposed, and some suggestions are offered. Finally, this study will give the analysis and prospect of the future trend of plant diseases and pests detection based on deep learning.