Dynamic UAV Phenotyping for Rice Disease Resistance Analysis Based on Multisource Data.

Area of Science:

• Agricultural Science
• Plant Pathology
• Remote Sensing Technology

Background:

• Bacterial blight significantly threatens rice production and food security, necessitating resistant cultivars.
• Traditional disease resistance evaluation is time-consuming and laborious.
• Unmanned Aerial Vehicle (UAV) remote sensing offers a faster alternative for infield phenotype evaluation.

Purpose of the Study:

• To develop an effective model for evaluating rice bacterial blight severity using time series UAV remote sensing and accumulated temperature data.
• To assess the scalability and utility of the developed model across different geographical locations.
• To integrate high-throughput phenotyping with quantitative trait loci (QTL) analysis for accelerated breeding of disease-resistant rice.

Main Methods:

• Utilized time series UAV remote sensing data and accumulated temperature data to train a bacterial blight severity evaluation model.
• Employed a model updating strategy to test scalability across different geographical locations.
• Combined UAV-based phenotypic analysis with QTL analysis to identify resistance QTLs in genetic populations.

Main Results:

• The predictive model achieved an R_p^2 of 0.86 and an RMSE_p of 0.65.
• A model updating strategy using 20% of transferred data proved effective for evaluating disease severity across different sites.
• Identified three new resistance QTLs, noting that QTLs varied across different growth stages.

Conclusions:

• Time series UAV remote sensing data, combined with accumulated temperature, provides an effective method for evaluating rice bacterial blight severity.
• The developed model is scalable to different geographical locations with minimal data transfer.
• Integrating UAV high-throughput phenotyping with QTL analysis accelerates the identification of resistance genes, aiding in disease resistance breeding.