Research on tomato leaf disease identification based on deep learning

Abstract

Tomato is one of the important economic forest fruits in my country. It is the fourth largest vegetable and fruit in my country with an annual output of about 55 million tons, accounting for 7% of the total vegetables. Due to the wide planting area, large yield, and high-quality vegetables are the development direction of modern agriculture. Therefore, this paper adopts the deep learning method, uses the CNN to collect the leaves of tomato diseases and pest detection, uses the stacking to detect the diseases and insect pests on the leaves with the optimized DenseNet121 and MobileNet-V2, and compares the individual DenseNet121 model and MobileNet-V2 model. It shows that the detection results of pests and diseases after fusion are higher than other algorithms, and the final detection accuracy reaches 98.24%, which effectively improves the detection accuracy. It provides a more effective method for the treatment of tomato diseases and insect pests.