Intelligent Detection of Mango Disease Spores Based on Mask Scoring R-CNN

In the process of mango cultivation, there has been a major threat of mango diseases and insect pests. Mango diseases and insect pests are diverse, including mango anthracnose, mango black spot disease and Curvularia leaf spot disease, and anthrax spores, Alternaria alternata and Curvularia spp Source of infection. By detecting the presence of these three kinds of spores, the future disease situation of mango can be estimated, which is of great significance for the prevention and control of mango diseases and insect pests. In this paper, three different types of spores in the microscopic image were detected and segmented by means of deep learning, so as to obtain the types of spores and the number of various spores in the image, and then the research results were fused with the Web system to make a detection system for mango disease spores. Compared with the Mask R-CNN network, the Mask Scoring R-CNN network adopted in this paper has stronger performance, and the trained model has better segmentation effect. Deploying the model to the Web system enables everyone to detect the type of spores and count the spores anytime and anywhere, which enhances the practicability of the intelligent detection work. The work done in this paper can play a positive role in the prevention and control of mango diseases.