Answers are grounded in this paper and cite their sources.
● Identifying the sensitive hyper-spectral bands for cotton VWS ● Evaluating sensitive 2D VI using correlation analysis ● Comparing different machine learning models for detecting VWS ● RF coupled with ratio index outperformed than common models ● Data updating strategy tested the model stability Verticillium Wilt (VW) is a common soilborne disease of cotton, and this disease has substantially reduced the cotton yield. Therefore, timely and accurate assessment of cotton VW Severity (VWS) is crucial for disease control and yield protection. In this research, in-situ investigations were conducted across 3 experimental sites and 2 years 2019 and 2021, where VWS values and spectral data were collected from a total of 361 cotton canopies. To construct models for monitoring VWS of cotton, two-band (2D) spectral index analysis approach was employed to identify the sensitive band and Vegetation Indices (VI). Following this, datasets in form of full spectra, sensitive band, and 2D VIs were separately integrated with machine learning algorithms: Random Forest (RF), Support Vector Machine (SVM), Back Propagation Neural Network (BPNN), and Convolutional Neural Network (CNN) to develop the monitoring models. A data-updating strategy was adopted to evaluate model stability by progressively enlarging the training pool. The training subset was initially set to 10% of the total data and expanded in 10% increments to 90%, with the validation subset reduced symmetrically at each step. Models were retrained and retested at every increment. The results showed that the top 4 models with data were: RF model with Refined Ratio Vegetation Index (RRVI) (R 2 =0.83, RMSE=8.09), RF model with Refined Normalized Difference Vegetation Index (RNDVI) (R 2 =0.82, RMSE=8.26), and BPNN model with RRVI (R 2 =0.79, RMSE=9.00), and RF model with sensitive band (R 2 =0.78, RMSE=9.78), respectively. The data updating strategy demonstrated that accuracy stabilized once 40% of the total data had been incorporated in this current study. The proposed data updating strategy with machine learning models offered new possibilities for the rapid and accurate assessment of cotton VWS.
No takes yet. Share an insight, caveat, or question.
Guo et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: