Method for structuring wheat blumeriagraminis speer early detection model by extracting sensitive parameters on basis of subwindow permutation analysis

Abstract

The invention provides a method for structuring a wheat blumeriagraminis speer early detection model by extracting sensitive parameters on the basis of subwindow permutation analysis (SPA). The method comprises the following steps of 1) acquiring the hyperspectral reflectance of an infected wheat; 2) extracting a sensitive wave band from the original wave band of the hyperspectral reflectance through the SPA algorithm; 3) selecting spectral indexes possibly related to diseases in existing research, and extracting sensitive spectral indexes from the spectral indexes through the SPA algorithm; 4) through partial least square-linear discriminant analysis, taking the sensitive wave band or the sensitive spectral indexes as component input variables to structure the wheat blumeriagraminis speer early detection model; 5) testing the wheat blumeriagraminis speer early detection model through the dichotomy algorithm, evaluating model performance through leave-one-out cross validation on the basis of independent susceptible varieties. Therefore, a wheat blumeriagraminis speer sensitive spectrum extracted on the basis of SPA is accurate in characteristic and few in wave bands, and the structured wheat blumeriagraminis speer detection model is simple, high in accuracy and good in stability.

Description

technical field [0001] The invention relates to the technical field of early prediction of plant diseases, in particular to a method for constructing an early monitoring model of wheat leaf powdery mildew by extracting sensitive parameters based on a sub-window rearrangement method. Background technique [0002] Powdery mildew (Blumeria graminis Speer) is a major worldwide disease in wheat production, and it is also one of the main diseases affecting wheat yield. Early identification, rapid monitoring, and quantitative evaluation of wheat powdery mildew are the core key technologies for precise breeding of wheat powdery mildew resistance, precise pesticide application, ecological security and loss assessment. The previous monitoring of crop diseases was mainly through destructive sampling measurement or field investigation of diseased plants to calculate the incidence, severity or disease index, which was often time-consuming, laborious, inefficient, highly subjective, and p...

AI Technical Summary

Problems solved by technology

Although the variable importance evaluation method based on rearrangement is highly evaluated in the random forest algorithm, each variable in the random forest is only randomly rearranged once, so the obtained value cannot be repeated; the existence of random forest model interference variables has a masking effect, It may not be possible to filter out the optimal variable set

Images