abstract |
The technical scheme of the invention discloses a method and system for identifying diseased leaves of various types of citrus based on machine learning and hyperspectral imaging, and obtains hyperspectral information of five types of diseased leaves of citrus, including normal leaves, canker leaves, herbicide damaged leaves, red leaves Spiders damage leaves and leaves with sooty disease; the spectral information is used as the experimental sample, the experimental sample is preprocessed and the characteristic wavelength is extracted, and the support vector machine and random forest algorithm are used to design the diseased leaf identification model to realize the classification and identification of citrus diseased leaves. The advantage of the invention is that it combines hyperspectral imaging and machine learning technology to classify and identify various types of citrus diseased leaves, obtains the best classification model of five types of citrus diseased leaves, and provides an effective nondestructive detection method for citrus growth status monitoring and disease and insect pest identification. . |