Enhancing plant pathology discovery and application development through automated, high-throughput hyperspectral...

Area of Science:

• Plant pathology
• Spectroscopy
• Agricultural science

Background:

• Hyperspectral sensing offers powerful insights in plant pathology.
• Current data collection methods face bottlenecks in throughput, data handling, and volume.
• These limitations hinder the transition of hyperspectral technology from research to practical applications.

Purpose of the Study:

• To develop an automated high-throughput hyperspectral imaging (AHHI) platform.
• To overcome limitations in current hyperspectral data collection for plant pathology.
• To enable scalable, high-volume data acquisition for advanced applications.

Main Methods:

• An automated system integrating a push broom hyperspectral camera (400-1000 nm) and robotic sample positioning.
• Acquisition of line images at 100 frames per second, processing large data volumes (2.5 GB per sample, 9 TB per day).
• Demonstration of use cases including pre-symptomatic disease detection, fungicide detection, and grapevine lineage discrimination.

Main Results:

• The AHHI platform successfully collected high-quality hyperspectral images at unprecedented scale (9 TB/day).
• PERMANOVA and Random Forest analyses showed high accuracy (AUC 77.8-99.9%) for disease detection and lineage discrimination.
• Results mirrored handheld spectrometer accuracies, demonstrating the system's efficacy despite spectral resolution differences.

Conclusions:

• The AHHI platform significantly enhances throughput and data volume for hyperspectral plant pathology studies.
• This technology facilitates the translation of hyperspectral discoveries from niche research to widespread practical use.
• AHHI accelerates the adoption of advanced hyperspectral applications in agriculture and plant science.