A mulberry brown spot disease recognition method based on remote sensing technology and pathological feature fusion

Abstract

The application discloses a mulberry brown spot disease recognition method based on remote sensing technology and pathological feature fusion, relates to the technical field of agricultural information technology, and comprises the following steps: acquiring multispectral or hyperspectral remote sensing images of a target mulberry planting area at at least two different time points, pre-processing the remote sensing images, extracting reflectivity data of a mulberry canopy layer from the pre-processed images, calculating an average brown spot index BI_mean and an average green-red difference index GRDI_mean of the canopy layer region, extracting image features based on near-infrared band images, calculating a time change rate based on data of the at least two time points, combining into a feature vector, inputting the feature vector into a pre-trained support vector machine classifier, and outputting a disease grade. The application converts the brown spot disease pathological features into calculable image features, combines special spectral indexes to construct a specific recognition model, significantly improves accuracy and reduces false positives. Through multi-temporal data analysis of disease feature changes, dynamic monitoring and early warning are realized.

Images