A new 3D vision-based leaf rolling index (LRI) and its application as a stable indicator of cotton drought stress

Abstract

The leaf rolling index (LRI) is a phenotype with significant physiological implications under drought stress. However, research on the quantification of the cotton LRI is lacking, limiting its application in drought diagnosis, irrigation guidance, and physiological assessments. This study conducted a 3D reconstruction of cotton using Structure from Motion (SFM) and Multi-View Stereo (MVS). Algorithms for leaf point cloud preprocessing and phenotype extraction were developed using the PCL point cloud library and integrated into software to calculate the leaf area and perimeter. The LRI was quantified in 3D space based on the point cloud area ratio. On this basis, we analyze the relationships between LRI and leaf physiological indicators such as leaf water potential (LWP), relative water content (RWC), stomatal conductance (gs), and electron transport rate (ETR) at the seedling and flowering stages. The results indicate that the cotton LRI provides a stable indicator of drought stress, which is mainly reflected in the stable correlation between the LRI and water physiological parameters (LWP, and RWC), with coefficients of determination (R²) exceeding 0.70. Furthermore, the correlation between the LRI and the ETR suggests that the LRI could be used to assess photosynthetic efficiency under drought stress. This study demonstrates that LRI based on 3D vision in cotton may serve as a reliable morphological indicator for indicating drought stress and evaluating photosynthetic efficiency.

Keywords

• Crop phenotype
• 3D reconstruction
• Water diagnosis
• Photosynthetic metabolism