Identification And Analysis Of Pear Disease Based On Deep Learning

With the advent of the intelligent age,“artificial intelligence” has become an indispensable topic in people’s daily lives.In recent years,more domestic and foreign industry researchers have focused more on their development in the agricultural sector.With the introduction of computer vision technology,the term “neural network” has gradually entered people’s field of vision.The use of deep learning technology to accurately and quickly identify plant leaf diseases can increase crop yield,prevent plant diseases and insect pests,and effectively inhibit the harm of pesticides to agricultural products and the surrounding environment.In this paper,the pear tree disease is taken as the entry point,and the performance of mainstream deep learning technology in identifying pear tree brown spot disease and pear rust is analyzed.The main contents are as follows:(1)Image enhancement,noise elimination and feature extraction of pear tree leaves by image preprocessing.(2)On the basis of(1),training the segmented pear tree disease using support vector machine.(3)Combining VGGNet and Res Net two deep convolutional neural network models to train and identify diseased leaves.(4)In addition,the severity of plant diseases is graded,and the severity is divided into six levels,which are divided by the area of diseased leaves.After testing,the HSV color space can effectively extract the brown spot disease and pear rust lesions of pear trees.The severity of pear disease was also divided by counting the number of pixels.Moreover,the experiment also confirmed the advantages of VGGNet and Res Net two deep convolutional neural network models in image recognition tasks.At the same time,due to the reliability and adaptability of the results,the experimental results are of great help for the subsequent related disease identification.