A rapid detection method for plant diseases based on hyperspectral

Through the hyperspectral plant disease detection model, using spectral data for feature selection and model construction, the problems of time-consuming, labor-intensive and cost-effective detection in the existing technology are solved, and the rapid, accurate and non-destructive detection of plant diseases are achieved.

CN115326720BActive Publication Date: 2025-03-25HUAZHONG AGRI UNIV
2 Cites 0 Cited by

Patent Information

Application Number
CN202210924576.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-08-02
Publication Date
2025-03-25
Estimated Expiration
2042-08-02

AI Technical Summary

Technical Problem

The existing plant disease detection methods are time-consuming and laborious inquiry and difficult to detect pathogen invasion in the early stage. PCR detection is costly and inefficient, making it difficult to apply on a large scale.

Method used

A hyperspectral-based plant disease detection model is adopted, and the reflectance spectral data of plant samples are collected, and pre-processed, spectral enhancement, specific spectral feature selection, optimized feature combination and support vector machine modeling are carried out to achieve rapid lossless recognition of plant diseases.

Benefits of technology

It has achieved rapid, efficient, non-destructive, repeatable and high accuracy detection of plant diseases, and can detect diseases in a timely manner in the asymptomatic stage, and take early prevention and control measures to curb the spread and outbreak of diseases.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses a method for constructing a plant disease detection model based on a hyperspectral spectrum, comprising the following steps: 1) collecting reflectance spectral data of plant samples; 2) calibrating the collected data by spectral standardization, and smoothing the spectrum by Savitzky‑Golay filtering; 3) performing spectral enhancement processing on the reflectance spectral data processed in step 2) by spectral index calculation, spectral first-order derivative calculation, and continuous wavelet transform; 4) determining specific spectral features by using a RELIEF‑F algorithm and an overlay strategy; 5) determining the optimal spectral feature combination under different spectral enhancement processing based on a sequence floating forward selection algorithm; 6) combining multiple types of spectral enhancements to optimize multiple spectral enhancement features; 7) constructing a plant disease detection model by using a support vector machine algorithm. The invention also provides a rapid and non-destructive plant disease detection method.
Need to check novelty before this filing date? Find Prior Art