Previsual and Early Detection of Myrtle Rust on Rose Apple Using Indices Derived from Thermal Imagery and...

Area of Science:

• Plant pathology
• Agricultural remote sensing
• Fungal disease detection

Background:

• Myrtle rust (Austropuccinia psidii) poses a significant threat to Myrtaceae species, impacting commercial nurseries.
• Early and previsual detection of myrtle rust is crucial for effective disease management in propagation settings.

Purpose of the Study:

• To evaluate the accuracy of thermal indices and narrowband hyperspectral indices (NBHI) for previsual and early detection of myrtle rust.
• To identify key thermal and NBHI for distinguishing infected from healthy plants.

Main Methods:

• Time-series thermal imagery and visible-to-short-infrared spectroscopy were used on rose apple plants (Syzygium jambos).
• Data were collected over 10 days from both healthy and rust-inoculated plants.
• Models were developed using thermal and NBHI from green and red leaves.

Main Results:

• Thermal imagery models achieved perfect accuracy (100%) in distinguishing infected plants one day before visible symptoms (1 DBS).
• Myrtle rust infection led to lower, more variable normalized canopy temperature, linked to increased stomatal conductance and transpiration.
• NBHI from green leaves provided excellent previsual classification from 1 to 3 DBS (F1 score range: 0.89–0.94).

Conclusions:

• Thermal imagery and NBHI offer robust methods for previsual and early detection of myrtle rust.
• These techniques can be integrated into nursery management for timely disease control.
• Further development could lead to a reliable detection methodology for commercial nurseries.