Integration of X-Ray CT, Sensor Fusion, and Machine Learning for Advanced Modeling of Preharvest Apple Growth...

Area of Science:

• Agricultural Science
• Plant Physiology
• Imaging Technology

Background:

• Understanding fruit quality requires linking internal structure to environmental conditions.
• Traditional methods often focus on postharvest analysis, limiting insights into developmental processes.
• Non-destructive characterization of fruit internal architecture is crucial for advancing agricultural science.

Purpose of the Study:

• To introduce a novel application of dual-resolution X-ray computed tomography (CT) for non-destructive apple internal tissue analysis.
• To correlate apple internal tissue architecture with environmental factors influencing fruit quality.
• To develop predictive models for fruit quality based on structural and environmental data.

Main Methods:

• Dual-resolution X-ray computed tomography (CT) for non-destructive internal tissue imaging.
• Extraction of 3D structural parameters (e.g., porosity, heterogeneity).
• Statistical analyses including correlation, PCA, CCA, SEM, and machine learning (ML) models (MLR, Random Forest, XGBoost).

Main Results:

• Identified temperature as the primary environmental driver of fruit quality, particularly at 45 Days After Full Bloom (DAFB).
• Revealed nonlinear, hierarchical effects of vapor pressure deficit, relative humidity, and light on quality traits.
• Achieved high predictive accuracy (R² > 0.99 for MLR) with temperature as the key predictor.

Conclusions:

• Dual-resolution X-ray CT provides valuable insights into fruit internal structure and its relationship with environmental factors.
• Early-stage environmental conditions significantly influence apple quality development.
• This multidisciplinary framework supports precision orchard management by enhancing predictive precision for fruit quality.