Disease and pest identification method based on attention high-order residual network

Abstract

The invention discloses a pest and disease identification method based on an attention high-order residual network, and the method mainly comprises the following steps: preprocessing a crop pest and disease picture, and carrying out a series of preprocessing on the obtained pest and disease picture; feature extraction: adding an attention mechanism in a residual module, cascading the improved residual module, constructing a high-order residual network model, and obtaining crop disease and insect pest features; disease identification and classification: after passing through an average pooling layer and a Softmax layer, realizing classification and identification of the disease and pest image; the crop disease and insect pest identification method mainly aims at the problems of low crop identification rate and coarse-grained disease classification under a complex background, provides a method capable of fully extracting crop disease characteristics and classifying crop disease degrees, has relatively good anti-interference capability and strong robustness, and is suitable for popularization and application. And actual crop disease identification application requirements can be met.

Description

technical field [0001] The invention belongs to the field of image recognition, in particular to the recognition of fine-grained crop diseases and insect pests under complex backgrounds, and is applicable to the field of recognition of crop diseases and insect pests. Background technique [0002] my country is a large agricultural country. Crop diseases and insect pests are one of the main reasons for the decline of crop quality and economic losses of farmers, and are closely related to daily economic activities. In traditional agricultural production, farmers use experience to identify and diagnose crop diseases, which cannot guarantee the accuracy and reliability of diagnosis. Since agricultural experts cannot provide real-time on-the-spot guidance in the field, it is extremely important to accurately identify crop diseases with the help of information technology. When most crop diseases occur, diseased spots appear on the leaves of crops, with different color, shape and ...

AI Technical Summary

Problems solved by technology

Research at home and abroad shows that this kind of method has achieved good recognition results, but it has certain limitations when it comes to the practical application of crop disease recognition: a large amount of image preprocessing is time-consuming and labor-intensive; features based on artificial design ignore The global and contextual information in the diseased image can easily lead to the problem of semantic gap

Therefore, the problem of performing fine-grained disease identification remains to be solved

Images