Overcoming field variability: unsupervised domain adaptation for enhanced crop-weed recognition in diverse farmlands

Frontiers in Plant Science

Plant Science

Jonghoon Lee¹  Hyongsuk Kim²  Yongchae Jeong²  Talha Ilyas³  Okjae Won⁴ 

[1]Core Research Institute of Intelligent Robots, Jeonbuk National University, Jeonju-si, Republic of Korea

[2]Core Research Institute of Intelligent Robots, Jeonbuk National University, Jeonju-si, Republic of Korea

[3]Division of Electronics Engineering, Jeonbuk National University, Jeonju-si, Republic of Korea

[4]Division of Electronics and Information Engineering, Jeonbuk National University, Jeonju-si, Republic of Korea

[5]Core Research Institute of Intelligent Robots, Jeonbuk National University, Jeonju-si, Republic of Korea

[6]Production Technology Research Division, National Institute of Crop Science, Rural Development Administration, Miryang, Republic of Korea

关键词: crop-weed recognition;    domain adaptation;    precision agriculture;    artificial intelligence;    crop phenotyping;    agricultural operations;   

DOI  :  10.3389/fpls.2023.1234616

 received in 2023-06-05, accepted in 2023-07-06,  发布年份 2023

来源: Frontiers

【 摘 要 】

Recent developments in deep learning-based automatic weeding systems have shown promise for unmanned weed eradication. However, accurately distinguishing between crops and weeds in varying field conditions remains a challenge for these systems, as performance deteriorates when applied to new or different fields due to insignificant changes in low-level statistics and a significant gap between training and test data distributions. In this study, we propose an approach based on unsupervised domain adaptation to improve crop-weed recognition in new, unseen fields. Our system addresses this issue by learning to ignore insignificant changes in low-level statistics that cause a decline in performance when applied to new data. The proposed network includes a segmentation module that produces segmentation maps using labeled (training field) data while also minimizing entropy using unlabeled (test field) data simultaneously, and a discriminator module that maximizes the confusion between extracted features from the training and test farm samples. This module uses adversarial optimization to make the segmentation network invariant to changes in the field environment. We evaluated the proposed approach on four different unseen (test) fields and found consistent improvements in performance. These results suggest that the proposed approach can effectively handle changes in new field environments during real field inference.

【 授权许可】

Unknown   
Copyright © 2023 Ilyas, Lee, Won, Jeong and Kim