An Open-Source Package for Thermal and Multispectral Image Analysis for Plants in Glasshouse.

Area of Science:

• Agricultural Science
• Plant Biology
• Computational Biology

Background:

• Advanced plant phenotyping accelerates the development of improved crop varieties.
• Accurate measurement of biophysical traits is crucial for crop breeding and stress management.
• Existing phenotyping methods require optimization for image acquisition and analysis.

Purpose of the Study:

• To develop a computational pipeline for synchronizing thermal and multispectral image acquisition.
• To optimize image pre-processing and analysis for accurate plant trait measurement.
• To provide a digital platform for monitoring crop stress and screening genotypes.

Main Methods:

• Developed a computational pipeline to correct for optical distortion, image co-registration, radiometric rescaling, and illumination.
• Synchronized image acquisition from thermal and multispectral sensors.
• Created a processed stacked image containing RGB, green, red, Near-Infrared (NIR), red edge, and thermal data, focusing on plant canopy pixels.

Main Results:

• The pipeline successfully synchronizes image acquisition from multiple sensors.
• Processed images provide accurate, object-specific data for analysis.
• Multimodal thermal and multispectral data enable complementary insights and effective development of vegetation indices.

Conclusions:

• The developed pipeline offers a robust solution for optimizing plant phenotyping with thermal and multispectral imagery.
• This digital platform and analytics facilitate early stress detection and high-throughput genotype screening.
• The open-source nature of the pipeline promotes wider adoption and advancement in crop phenotyping research.