Integration of temperature control and imaging-based prediction for grafting-related morphological regulation in tomato seedlings in a closed transplant production system

Tomato seedling growth and quality are crucial determinants of the success of grafting and transplant establishment. This study aimed to investigate temperature control strategies in a closed transplant production system (CTPS) and their integration with imaging-based prediction to produce grafting-ready seedlings. Scion ‘Dotaerang Dia’ and rootstock ‘B-Blocking’ were grown under combinations of average daily temperatures (ADTs; 24 and 26 °C) and difference between day and night temperatures (DIFs; –8, –4, 0, +4, and +8 °C). Morphological traits crucial for grafting, including the epicotyl length (EPL) and diameter (EPD) of scions and hypocotyl length (HYL) and diameter (HYD) of rootstocks, and canopy traits, including leaf area index (LAI) and canopy height (CH), were evaluated. Higher ADTs and positive DIFs promoted elongation, whereas lower ADTs and negative DIFs restricted elongation and improved compactness. Compact seedlings with a higher dry matter content are advantageous for grafting, whereas seedlings with greater elongation and dimensional synchrony better meet the requirements of robotic grafting. Imaging-based monitoring using multispectral-derived LAI and light detection and ranging (LiDAR)-derived CH accurately predicted grafting-related traits (R2 > 0.9 for EPL, EPD, and HYD); however, HYL predictions were less reliable under negative DIFs. Leave-one-environment-out cross-validation confirmed robust performance for diameter traits across environments. These findings collectively indicate that CTPS enable precise morphological regulation of tomato scions and rootstocks through temperature control, whereas imaging-based phenotyping allows a basis for real-time prediction of grafting suitability. This integration establishes a scalable and automation-ready framework for grafted transplant production, offering a technological foundation for developing automated grafting strategies.