I developed a ResNet-based model to predict plant root depth and diameter using data collected from in-soil fiber Bragg grating (FBG) sensors. We used simulated data as well as  corn root data. This work was accepted for publication in IEEE Photonics Technology Letters (PTL) on March 17, 2025 (Link), and another paper is currently in preparation. To scale up, we plan to implement distributed fiber sensors to monitor multiple plants simultaneously.

I developed a Fusarium wilt detection model using ResNet to analyze drone images in collaboration with the Yuma Agricultural Center. The model achieved 96.30% accuracy and 97.10% F1-score, with an 86% accuracy in quantifying disease severity. A journal paper has been submitted to Artificial Intelligence in Agriculture journal, and I plan to extend this approach to other crops to improve disease detection and management across diverse agricultural systems. 

This project focuses on understanding root-root interactions, which significantly influence architectural root phenotypes. Comprehending Root System Architecture (RSA) is essential for improving nutrient and water uptake, enhancing crop resilience, and contributing to food security. I collaborated with a PhD student for the project, mentoring the student and developing plant root segmentation models using U-Net, which serves as the first step in understanding root-root interactions. 

In my research, I analyze the impact of weather parameters on Fusarium severity using five years of weather data (2019-2023) in collaboration with Yuma Agricultural center. Utilizing statistical models, including mixed models and ordinary least squares (OLS), I examine how factors like air temperature, humidity and soil temperature influence Fusarium severity in crops. This is an ongoing project.