Crop disease identification method based on incremental learning

Abstract

The invention discloses a crop disease identification method based on incremental learning, and relates to the technical field of crop cultivation, and the method comprises the following steps: building a convolutional neural network model for identifying disease image features through pre-collecting a target region for demarcating leaf diseases, converting a preprocessed image into an HSI space image, and obtaining an HSI space image; extracting color features and texture features of the images, determining an Euclidean distance between each to-be-detected image and the disease training set image under each feature, correcting the extracted significant identification area to obtain disease spot images, and inputting the obtained disease spot images into an identification convolutional neural network model for identification and classification. According to the method, rotation invariance and illumination invariance of the extracted color features and texture features are achieved, detection and positioning of crop disease parts in a real environment and accurate identification of disease types are achieved, the method is suitable for prevention, control and early warning of crop diseases, and workers can take targeted disease prevention and control measures conveniently.

Description

technical field [0001] The invention relates to the technical field of crop cultivation, in particular to a method for identifying crop diseases based on incremental learning. Background technique [0002] Crop disease is one of the main agricultural disasters in my country. It has the characteristics of various types, great impact and frequent outbreaks. It not only causes losses to crop production, but also poses a threat to food safety. Therefore, the diagnosis and identification of crop diseases play an important role in ensuring crop yield and preventing food safety. At the same time, realizing accurate detection of crop diseases and the determination of the degree of disease is the key to the prevention and control of crop diseases. At present, the traditional crop disease identification mainly relies on the experience accumulated by farmers in the agricultural production process to make judgments, which is time-consuming and labor-intensive, and the real-time and accu...

AI Technical Summary

Problems solved by technology

At the same time, in the feature extraction, the texture features in the existing methods are mainly correlation, energy, entropy, contrast and deficit, etc., which define many texture features based on the statistical method based on the gray level co-occurrence matrix. Compared with global features, local features have significant advantages such as rotation invariance and gray invariance; and the identification and classification of diseases in crop disease identification and classification mainly use neural networks, support vector machines and improved support vector machine methods, although these The method can identify the types of diseases, but the number of types identified by these methods is small, generally only three types of diseases are identified, and these methods require a lot of samples for training, and the recognition rate is not high

Images