Crop disease and pest detection and prevention device and method based on image and deep learning

Abstract

The invention discloses a crop disease and pest detection and prevention method based on images and deep learning, and the method comprises the steps: training a deep learning model through a master control device, and obtaining the image features of crop diseases and pests; further comparing the binarized image with the noise interference information eliminated with the image features of crop diseases and insect pests, and outputting a disease and insect pest classification result and the two-dimensional position coordinates (x, y) of the tea leaves with diseases and insect pests. The network structure for training the deep learning model is designed and built for crop diseases and insect pests, the network structure is simple, the training cost is low, and the detection accuracy is high.

Description

technical field [0001] The invention relates to the technical field of artificial intelligence and integrated control, in particular to a device and method for detecting and preventing crop diseases and insect pests based on images and deep learning. Background technique [0002] At present, thanks to the high degree of cooperation between the two technologies of machine vision and deep learning, practical applications have been launched in many engineering fields, especially in the fields of industrial sorting and security identification. Substantial progress has been made. [0003] However, in the field of agricultural equipment, the level of mechanical automation is relatively high, but the level of intelligence is far from enough. If the actual harvesting, pest identification and other processes are still dominated by manpower, the labor efficiency is low, or there is no difference in comprehensive sprinkler irrigation and control. Also can cause the excessive and wastef...

AI Technical Summary

Problems solved by technology

[0003] However, in the field of agricultural equipment, the level of mechanical automation is relatively high, but the level of intelligence is far from enough. If the actual harvesting, pest identification and other processes are still dominated by manpower, the labor efficiency is low, or there is no difference in comprehensive sprinkler irrigation and control. It will also cause the problem of excessive and wasteful pesticides

Even though machine vision technology has been initially applied in the field of agricultural equipment, there are problems such as low recognition accuracy and poor ability to resist environmental interference. Therefore, how to improve the recognition accuracy of machine vision in the actual application environment has become an inevitable engineering problem.

[0004] In recent years, with the wide application of deep learning, it has begun to exert its strength in various industries, such as the Yolo series model and Faster R-CNN model, which have begun to be applied to the field of agricultural equipment, but the existing training model structure is relatively redundant, and Poor real-time performance

Taking the Yolo-v4 model as an example, there are a total of 53 layers of convolutional networks, which are more suitable for classification and detection of dozens of objects, and because of the high complexity of the model, the cost of training time is high. In addition, the existing technology uses Deep learning technology to detect the types of pests and diseases mostly stays in the detection and identification link. There is no integrated solution for how to prevent and control the problem

Images