Crop disease and pest identification method based on improved VGG-16 network

Abstract

The invention discloses a crop disease and pest identification method based on an improved VGG-16 network, and the method comprises the steps: S1, obtaining crop disease and pest image data, and carrying out the classification according to different crop diseases and pests; s2, inputting an image in the crop disease and insect pest image data into an improved VGG-16 network for parameter training; and S3, after parameter training is completed, obtaining an optimal network capable of identifying crop diseases and insect pests in the image, and finally automatically identifying the crop diseases and insect pests in any input image through the optimal network. According to the method, the problems of low accuracy, lack of robustness, poor adaptive capacity and the like in the prior art are solved, the robustness is higher, the recognition rate is higher, network parameters are fewer, the adaptive capacity is higher, the recognition speed is higher, and the crop diseases and pests in the image are recognized more accurately.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a crop disease and insect pest recognition method based on an improved VGG-16 network. Background technique [0002] With the rapid development of my country's agriculture, the requirements for the yield and quality of crops are also increasing. Crop diseases and insect pests are one of the important factors affecting the yield and quality of crops. In the past, manual identification was used, that is, technicians were used to identify crop diseases and insect pests. However, the level of knowledge and experience of technicians was uneven, resulting in high uncertainty of identification accuracy, and it could only deal with small-scale planting. In this case, there are disadvantages such as low efficiency, poor real-time performance and large cost of manpower and material resources. [0003] Smart agriculture requires the realization of unmanned, automated, and intell...

AI Technical Summary

Problems solved by technology

In the past, manual identification was used, that is, technicians were used to identify crop diseases and insect pests. However, the level of knowledge and experience of technicians was uneven, resulting in high uncertainty of identification accuracy, and it could only deal with small-scale planting. In some cases, there are disadvantages such as low efficiency, poor real-time performance and large cost of manpower and material resources.

[0003] Smart agriculture requires the realization of unmanned, automated, and intelligent agricultural management. Many farms obtain crop images by installing high-definition cameras, and use traditional visual learning methods to analyze the images obtained to identify crop diseases and insect pests, which can solve the shortcomings of low efficiency of manual identification, and also meet the needs of smart agriculture. oriented, but there are still disadvantages such as low accuracy, lack of robustness and poor adaptability

Images