Construction of wheat rust disease inversion model based on multi-spectral image features

Journal Title: Journal of Henan Agricultural University - Year 2023, Vol 57, Issue 5

Abstract

[Objective] This paper studied how to extract effective features from multi-spectral remote sensing images of unmanned Aerial Vehicle and build wheat rust monitoring model. [Method] This study takes wheat rust fields as the research object, and uses the Dji Spirit 4 drone to obtain multispectral images. The vegetation index and texture index are extracted, and Pearson correlation coefficient, gray correlation degree, and variable projection importance are used for index sorting. The decision coefficient and Akihchi information criterion are used for variable screening, and three methods, partial least squares regression (PLSR), back propagation neural network (BPNN), and random forest (RF), were respectively used to establish wheat rust inversion regression model. [Result] The experimental results show that the VIP-AIC can screen out good spectral indices. The regression model constructed using the BPNN method has the highest accuracy, with a determination coefficient R2 of 0.918 and RMSE of 0.128. [Conclusion] Employed three feature selection methods to identify sensitive features from the constructed set of spectral indices and texture features. Subsequently, a wheat rust severity inversion model was established, enabling the mapping of the spatial distribution of wheat rust disease. These findings provide a methodological reference for feature variable selection in multispectral image analysis.

Authors and Affiliations

Zhiye WANG, Tao LIU, Huan ZHANG, Quanguo ZHANG, Haitao ZHANG, Yuhong HE

Keywords

Related Articles

Research progress on main active components and multifunctional utilization of aromatic plants

In this paper, the main active components of common aromatic plants were analyzed, the aroma characteristics of terpenoids, aromatic family of benzene ring structure and fatty acids in aromatic components, the medicinal...

Cloning, expression and functional analysis of LlARF12 gene in Lilium lancifolium

[Objective] In order to explore the internal mechanism of bulbil formation in Lilium lancifolium, the cellular and histological changes before bulbil formation were discussed. The auxin response factor LlARF12 was cloned...

A study on the bone characteristics of local chickens in Henan Province

[Objective] This paper aims to study the characteristics and differences of skeletal characteristics of local chickens in Henan Province. [Method] Adult chickens of five local chicken breeds in Henan Province, including...

Spatio-temporal heterogeneity of planting carbon footprint in Yellow River Basin under the carbon peaking and carbon neutrality targets

[Objective] This study was conducted to explore the spatial distribution characteristics of planting carbon footprint in the Yellow River Basin and its spatio-temporal heterogeneity with various influencing factors. [Met...

Performance test and analysis of a photovoltaic pumping system for farmland irrigation

[Objective] This study was conducted to explore the comprehensive energy demand characteristics of photovoltaic water pump systems for farmland irrigation, and improve the reliability and practicality of the system's per...

Download PDF file
  • EP ID EP769254
  • DOI 10.16445/j.cnki.1000-2340.20230905.001
  • Views 16
  • Downloads 0

How To Cite

Zhiye WANG, Tao LIU, Huan ZHANG, Quanguo ZHANG, Haitao ZHANG, Yuhong HE (2023). Construction of wheat rust disease inversion model based on multi-spectral image features. Journal of Henan Agricultural University, 57(5), -. https://europub.co.uk/articles/-A-769254