AI-Assisted Phenotyping and Physiological Modeling for Yield Stability in Future Cropping Systems

https://doi.org/10.5281/zenodo.21506926

Abstract

Climate change has intensified the frequency of drought, heat, and other abiotic stresses, threatening global crop productivity and yield stability. Conventional breeding and process-based crop models have independently contributed to crop improvement; however, both remain constrained by limited phenotypic information, labor-intensive parameterization, and insufficient capacity to capture dynamic genotype × environment × management (G × E × M) interactions. Recent advances in high-throughput phenotyping (HTP), artificial intelligence (AI), machine learning (ML), and remote sensing have created new opportunities for integrating real-time phenotypic observations with physiological crop models. This review examines the convergence of AI-assisted phenotyping and process-based crop modeling as a framework for improving yield prediction and climate resilience. It discusses major sensing technologies, including RGB, hyperspectral, thermal, LiDAR, and root imaging systems, together with functional trait extraction methods. The review further summarizes the mathematical foundations of leading crop growth models such as DSSAT, APSIM, WOFOST, AquaCrop, and InfoCrop, followed by recent developments in data assimilation techniques, including Ensemble Kalman Filters, variational methods, and particle swarm optimization. Field-based case studies in wheat, cotton, canola, and spinach demonstrate substantial improvements in crop monitoring and prediction accuracy through the assimilation of high-throughput phenomic data. Finally, the emerging Knowledge- and Data-Driven Modeling (KDDM) paradigm is explored, highlighting parallel, serial, and modular integration strategies that combine mechanistic understanding with data-driven intelligence. The synthesis demonstrates that AI-assisted phenotyping integrated with physiological modeling provides a robust pathway toward digital agriculture, precision crop management, accelerated breeding, and enhanced yield stability under future climate scenarios.