Sorting biotic and abiotic stresses on wild rocket by leaf-image hyperspectral data mining with an artificial...

Area of Science:

• Plant pathology
• Agricultural engineering
• Remote sensing

Background:

• Wild rocket (Diplotaxis tenuifolia) is susceptible to soil-borne stresses in intensive cultivation.
• Early stress detection is crucial for optimizing crop management and reducing external inputs.

Purpose of the Study:

• To investigate the potential of hyperspectral imaging for early detection of abiotic and biotic stresses in wild rocket.
• To identify specific spectral regions and wavelengths indicative of different stress types.

Main Methods:

• Hyperspectral image analysis of wild rocket plants under controlled stress conditions (water deficit, salinity, Fusarium, Rhizoctonia).
• Application of artificial intelligence (AI) models for image-based stress classification.
• Correlation of spectral data with physiological parameters.

Main Results:

• Water deficit and fungal diseases significantly impacted plant water status and pigment content.
• Biotic stresses induced visible discoloration within one week.
• Specific vegetation indices and AI-selected spectral bands (VIS and NIR) effectively discriminated between stress types.

Conclusions:

• Hyperspectral imaging provides valuable spectral information for identifying specific stresses in wild rocket.
• Narrowed spectral regions and single wavelengths can indicate changes in pigment and water status, revealing the stress etiology.