2024
Liu, David; Li, Zhengkun; Wu, Zihao; Li, Changying
Digital Twin/MARS-CycleGAN: Enhancing Sim-to-Real Crop/Row Detection for MARS Phenotyping Robot Using Synthetic Images Journal Article
In: Journal of Field Robotics, vol. n/a, no. n/a, 2024.
Abstract | Links | BibTeX | Tags: digital twin, field-based robotic phenotyping, object detection, sim-to-real transfer, zero-shot
@article{Liu,
title = {Digital Twin/MARS-CycleGAN: Enhancing Sim-to-Real Crop/Row Detection for MARS Phenotyping Robot Using Synthetic Images},
author = {David Liu and Zhengkun Li and Zihao Wu and Changying Li},
url = {https://onlinelibrary.wiley.com/doi/abs/10.1002/rob.22473},
doi = {https://doi.org/10.1002/rob.22473},
year = {2024},
date = {2024-12-10},
journal = {Journal of Field Robotics},
volume = {n/a},
number = {n/a},
abstract = {ABSTRACT Robotic crop phenotyping has emerged as a key technology for assessing crops' phenotypic traits at scale, which is essential for developing new crop varieties with the aim of increasing productivity and adapting to the changing climate. However, developing and deploying crop phenotyping robots faces many challenges, such as complex and variable crop shapes that complicate robotic object detection, dynamic and unstructured environments that confound robotic control, and real-time computing and managing big data that challenge robotic hardware/software. This work specifically addresses the first challenge by proposing a novel Digital Twin(DT)/MARS-CycleGAN model for image augmentation to improve our Modular Agricultural Robotic System (MARS)'s crop object detection from complex and variable backgrounds. The core idea is that in addition to the cycle consistency losses in the CycleGAN model, we designed and enforced a new DT/MARS loss in the deep learning model to penalize the inconsistency between real crop images captured by MARS and synthesized images generated by DT/MARS-CycleGAN. Therefore, the synthesized crop images closely mimic real images in terms of realism, and they are employed to fine-tune object detectors such as YOLOv8. Extensive experiments demonstrate that the new DT/MARS-CycleGAN framework significantly boosts crop/row detection performance for MARS, contributing to the field of robotic crop phenotyping. We release our code and data to the research community (https://github.com/UGA-BSAIL/DT-MARS-CycleGAN).},
keywords = {digital twin, field-based robotic phenotyping, object detection, sim-to-real transfer, zero-shot},
pubstate = {published},
tppubtype = {article}
}
ABSTRACT Robotic crop phenotyping has emerged as a key technology for assessing crops' phenotypic traits at scale, which is essential for developing new crop varieties with the aim of increasing productivity and adapting to the changing climate. However, developing and deploying crop phenotyping robots faces many challenges, such as complex and variable crop shapes that complicate robotic object detection, dynamic and unstructured environments that confound robotic control, and real-time computing and managing big data that challenge robotic hardware/software. This work specifically addresses the first challenge by proposing a novel Digital Twin(DT)/MARS-CycleGAN model for image augmentation to improve our Modular Agricultural Robotic System (MARS)'s crop object detection from complex and variable backgrounds. The core idea is that in addition to the cycle consistency losses in the CycleGAN model, we designed and enforced a new DT/MARS loss in the deep learning model to penalize the inconsistency between real crop images captured by MARS and synthesized images generated by DT/MARS-CycleGAN. Therefore, the synthesized crop images closely mimic real images in terms of realism, and they are employed to fine-tune object detectors such as YOLOv8. Extensive experiments demonstrate that the new DT/MARS-CycleGAN framework significantly boosts crop/row detection performance for MARS, contributing to the field of robotic crop phenotyping. We release our code and data to the research community (https://github.com/UGA-BSAIL/DT-MARS-CycleGAN).