An ai modeling method and system for evaluating cold resistance of pennisetum

Abstract

The application provides an AI modeling method and system for evaluating cold resistance of Pennisetum, and relates to the technical field of artificial intelligence, wherein the method comprises the following steps: acquiring multispectral image data of a Pennisetum sample plot, and acquiring real cold resistance labels of the Pennisetum corresponding to the multispectral image data in space; pre-processing the multispectral image data to extract pre-processed spectral feature data reflecting biological characteristics of the Pennisetum; taking the pre-processed spectral feature data as model input, taking the real cold resistance labels as model output, constructing and training an AI model until the AI model converges to obtain a trained model for evaluating cold resistance of the Pennisetum. The application discloses a deep correlation between spectral features and cold resistance phenotypes, thereby realizing early and accurate prediction and evaluation of cold resistance of the Pennisetum.

Images