Method for distinguishing rice blast disease in rice based on aerial photographs of farmland

Abstract

The invention discloses a method for distinguishing a rice blast disease in rice based on aerial photographs of a farmland. The method comprises the steps of image preprocessing, image segmentation, extraction and identification of diseased spots and calculation of damage percentage. In the step of image segmentation, color component combined images are segmented by adopting an Otsu method according to background of the images and diseased spot color characteristics; in the step of extraction and identification of the diseased spots, specially, the extraction and identification of the diseasedspots can be used for identifying that the diseased spots are the diseased spots of the rice blast disease or natural yellowing diseased spots; and finally, by combining with national grading standard, the grades of disease disasters of the rice blast disease are determined according to the area proportion of the diseased spots of the rice blast disease on the leaf blades. According to the methoddisclosed by the invention, acquired photographs with complicated backgrounds can be processed preferably, interference is reduced, the type of the diseased spots can be accurately distinguished, andthe disaster grade of the rice blast disease can be accurately determined.

Description

technical field [0001] The invention belongs to the field of monitoring by using image processing and analysis technology, and in particular relates to a method for discriminating rice blast diseases based on aerial farmland images. Background technique [0002] Rice blast is one of the three major diseases of rice, which seriously affects the production of rice. It not only reduces the yield of rice by 10-30%, but also reduces the quality of rice. At present, the classification of the disease degree of rice blast is mainly completed by manual visual inspection. This classification is highly subjective, requires relatively high professional quality of workers, and the classification efficiency is low, and there are still some errors. [0003] Ma Degui et al. used the ellipse model to detect the damage degree of rice blast and sheath blight, and the classification accuracy was 80% to 100%. Feng Lei et al. used multispectral imaging technology to extract rice leaf and canopy i...

AI Technical Summary

Problems solved by technology

At present, there are still relatively few studies on the classification of leaf blast diseases using image processing technology. At the same time, there are some shortcomings in the classification of diseases based on the area of ​​diseased spots to the area of ​​leaves. In the case of natural withered yellow leaves and infected withered yellow leaves, the accuracy of disease classification is not enough.

Images