Plant leaf disease identification method based on mask convolutional neural network

Abstract

The invention discloses a plant leaf disease identification method based on a mask convolutional neural network, and mainly solves the problem of low plant leaf disease identification accuracy in theprior art. According to the scheme, an original data set is enhanced and expanded to obtain a training set and a test set; performing semantic segmentation on the training set and the test set to obtain corresponding mask sets; adding a disease feature screening module between a full convolution layer and a mask branch of the model, and inputting the training set and the mask set into a network for training to obtain a target classification and target detection result; taking the feature map belonging to the diseased leaf in the target classification result as the input of a mask branch, and obtaining a trained model after multiple iterations; and inputting the test set into the model, carrying out target classification and target detection on the leaves, and segmenting the leaves belonging to the disease category. According to the method, the leaf disease identification accuracy is improved on the basis of a traditional method, and the method can be used for plant disease leaf identification and segmentation in agricultural planting.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a plant leaf disease identification method, which can be used for the segmentation and identification of plant disease leaves in agricultural planting. Background technique [0002] In modern smart agriculture, crop diseases are a major threat to food security, and plant diseases cause serious damage to crops by significantly reducing yields. Among them, early blight is a typical disease that can seriously reduce yield. Similarly, in humid climates, late blight is another very destructive disease that affects the leaves, stems and fruit of plants. Protecting plants from diseases is crucial to maintaining the quality and quantity of crops. Protecting crops should start with early detection of diseases so that appropriate treatments can be selected at the right time to prevent the spread of diseases. The types of diseases that identify greenhouse plant disea...

AI Technical Summary

Problems solved by technology

The images were collected from six different local datasets at different growth stages, lighting, resolution and soil types, but were less accurate and did not target plant disease detection and classification

[0005] In 2018, Liu Na and others used image processing technology and artificial neural network technology to realize the detection of cucumber leaf diseases and the classification of the degree of disease infection, and mainly for the high incidence and serious damage of cucumber downy mildew, powdery mildew and viral diseases. However, due to the small number of disease categories for identifying cucumbers, the number of training samples is small, and the number of disease types identified is small, the test accuracy is low and overfitting is prone to occur.

[0007] To sum up, the current research on plant diseases mainly focuses on the classification and identification of a single plant disease category similar to cucumber, which contains a small number of disease types and samples, and the identification accuracy is low.

The existing method for classifying plant diseased leaves through deep learning cannot segment the plant diseased leaves and the location of the disease, and the recognition rate is low

Images