A classification method of plant leaf diseases and insect pests based on transfer learning

Abstract

The invention discloses a method for classifying plant leaf diseases and insect pests based on migration learning. After the conventional image preprocessing, 0 filling processing is added, so that the edge feature information of the image can be extracted more clearly; the general machine vision provided by ImageNet is used The model weight parameters reduce the time complexity and resource consumption of model training; and improve the existing ResNet50 residual neural network, add an average pooling layer that can retain more information, and reduce multi-dimensional data to 1-dimensional Furthermore, the Flatten layer that speeds up the calculation speed, the first fully connected layer that calculates 2048 intermediate features, and the second fully connected layer at the end, compared with the existing technology, the training effect is better, and finally the processing effect with an accuracy of approximately 90% is obtained .

Description

technical field [0001] The invention relates to the field of agricultural machine vision, in particular to a method for classifying plant leaf diseases and insect pests based on transfer learning. Background technique [0002] With the development of agricultural modernization, the scale of planting area is increasing, and the problem of crop diseases and insect pests has become the primary problem. It has the characteristics of many types, great impact and frequent outbreaks, which often cause heavy losses to agricultural production. If you rely on professionals to observe crop diseases with the naked eye in the planting area, there are usually dozens of species that need to be checked, and the types of diseases of different species are also different. Not only is the labor cost huge, but the accuracy is not enough. [0003] In recent years, with the rapid development of machine vision, it has become an important research direction to solve the problem of pests and disease...

AI Technical Summary

Problems solved by technology

[0005] In view of the above-mentioned deficiencies in the prior art, a method for classifying plant leaf diseases and insect pests based on transfer learning provided by the present invention solves the problems of long model training time, large computational resource consumption and training accuracy of commonly used algorithms in the field of agricultural machine vision. optimization problem

Images