An image-based technique for automated root disease severity assessment using PlantCV

Area of Science:

• Plant pathology
• Computational biology
• Agricultural science

Background:

• Manual plant disease severity assessments are time-consuming and subjective.
• Traditional image analysis for root disease studies faces challenges like necrotic tissue removal.
• Automated methods are needed for fast, unbiased, and quantitative root disease severity measurements.

Purpose of the Study:

• To develop an automated Python-based pipeline (RootDS) for improved plant disease severity phenotyping.
• To generate binary images compatible with Root System Markup Language (RSML) software.
• To address limitations in current root disease assessment methodologies.

Main Methods:

• Developed the RootDS pipeline using PlantCV, a Python computer vision library.
• Applied the pipeline to common bean plants inoculated with Fusarium root rot.
• Generated quantitative disease scores and root area measurements.

Main Results:

• RootDS demonstrated strong correlations with manual assessments for disease scores (R² = 0.92) and root area (R² = 0.90).
• The pipeline captured a broader range of variation compared to manual scoring.
• Images generated by RootDS did not negatively impact RSML software output.

Conclusions:

• The RootDS pipeline offers enhanced functionality for plant disease score datasets.
• It provides a viable alternative for generating image sets for existing RSML software.
• Automated phenotyping improves consistency and quantitative data acquisition in plant pathology research.