Northeast black fungus disease and insect pest image recognition system based on deep learning

Abstract

The invention belongs to the technical field of image recognition systems, and particularly relates to a northeast black fungus disease and insect pest image recognition system based on deep learning.The system comprises a camera module, a network data transmission module, a computer terminal, a data processing module, a deep learning image recognition module, a classification result processing module and an output module. The camera module is in electrical output connection with the network data transmission module; the network data transmission module is in electrical output connection withthe computer end; the computer terminal is in electrical output connection with the data processing module; the data processing module is in electrical output connection with the network data transmission module; the network data transmission module is in electrical output connection with the classification result processing module; and the classification result processing module is in electricaloutput connection with the output module. By means of the agaric disease and insect pest detection system and method, a more complete agaric disease and insect pest data set can be obtained, the over-fitting degree of the model is effectively avoided, and the training model can have better robustness and good detection performance.

Description

technical field [0001] The invention relates to the technical field of image recognition systems, in particular to an image recognition system for Northeast black fungus pests and diseases based on deep learning. Background technique [0002] The traditional image classification technology mainly uses PC and embedded system to process and analyze the input image and select the features that can reflect the essence of the image, and finally divide the image into one of several categories according to the selected features. Among them, the selection and extraction of features and the selection of classifiers are very important in image classification. The traditional image classification algorithm relies entirely on the designer's subjective consciousness to extract the characteristics of the classification object for a specific classification scene. The manually selected image features generally include features such as color, texture, shape, and spatial relationship. The c...

AI Technical Summary

Problems solved by technology

[0004] During the large-scale cultivation of Northeast black fungus in greenhouses, the traditional prevention and control of diseases and insect pests adopts roughly the same method for each greenhouse, but due to the difference in the growth rate of the fungus and the differences in the surrounding environmental factors in each greenhouse, there are differences between the various greenhouses. The types of pests and diseases are also different, so if the same method is adopted, it will lead to low efficiency. At the same time, it is too time-consuming and labor-intensive to manually identify pests and diseases in each greenhouse

At the same time, traditional computer vision technology needs to manually design features. The quality of target detection results is directly related to the feature extraction algorithm. Therefore, the fitting degree of the model will fluctuate greatly due to the quality of the features taken. Different tasks may need to design Different feature extraction algorithms to achieve better recognition results

Images