Rice leaf blast detection and classification method based on multi-spectral image processing

Abstract

The invention discloses a rice leaf blast detection and classification method based on the multi-spectral image processing. A visible/near infrared multi-spectra camera is used for real-timely collecting monochromatic gray images of three waveband channels of green light wavebands, red light wavebands and near infrared wavebands, and then Matlab software is used for programming application software by an image processing method for image processing. The method comprises the removal of background, noises, interferences, and the like, and the recognition analysis of crop disease plague information, and the accurate and rapid processing of whether crops are ill and the position and the classification of disease plagues are realized. The disease recognition time of each picture is only a few seconds. The rice leaf blast detection and classification method based on the multi-spectral image processing is used for the rapid, accurate, stable, real-time and non-destructive crop leaf blast infection diagnosis and for accurately indicating the position of the disease plagues and the classification of infection grades, reducing the dosage due to the overall spraying, lowering the production cost, reducing pollution, providing data support for the variable spraying and improving the decision level of the accurate spraying. The method plays an active role in realizing the precision agriculture.

Description

technical field [0001] The invention relates to a method for detecting and grading rice leaf blast based on multispectral image processing. Background technique [0002] Rice blast is one of the most important diseases in rice production. It is widely distributed in various rice-growing areas in the world. It can occur during the entire growth period of rice, causing large-scale damage. In popular years, it generally reduces production by 10% to 20%. 40% to 50%. Among them, rice leaf blast and rice ear blast caused the greatest losses. The annual incidence rate of leaf blast is 40% to 50%; the heavy one can reach 80%; the ear blast can reach 50%. The population of blast fungus is complex and mutates rapidly, and blast-resistant varieties often lose their resistance after 3-5 years of popularization. However, in the current pesticide control, the disease data collection relies on the human eye for disease classification and reporting. The grading of leaf blast disease is ...

AI Technical Summary

Problems solved by technology

The grading of leaf blast disease is based on the appearance of lesions and the proportion of lesions to the area of ​​leaves. The classification standard requires more precise, and the ratio index of lesion damage area to leaf area between different disease levels is only 1%. It is difficult for people to accurately distinguish the disease

Even the analysis and judgment of experts with naked eyes takes a lot of time and energy, and the efficiency is not high

Images