Deep Learning Techniques for In-Crop Weed Identification: A Review
Kun Hu, Zhiyong Wang, Guy Coleman, Asher Bender, Tingting Yao, Shan Zeng, Dezhen Song, Arnold Schumann, Michael Walsh
TL;DR
The paper surveys deep learning techniques for image-based in-crop weed identification, addressing the core challenge of robust weed recognition under diverse field conditions. It systematically covers DL building blocks (from ML foundations to CNNs and GNNs) and maps them to four weed-recognition tasks: image classification, object detection, semantic segmentation, and instance segmentation, including representative architectures and public datasets. The review compiles insights from over 30 studies, discusses data modalities, augmentation, and evaluation metrics, and highlights practical challenges such as real-time inference, data scarcity, and deployment in large-scale farming. It concludes with a forward-looking set of opportunities, notably fine-grained learning, weakly-supervised/unsupervised methods, explainable AI, incremental learning, and the need for large, diverse datasets to enable field-ready weed control technologies.
Abstract
Weeds are a significant threat to the agricultural productivity and the environment. The increasing demand for sustainable agriculture has driven innovations in accurate weed control technologies aimed at reducing the reliance on herbicides. With the great success of deep learning in various vision tasks, many promising image-based weed detection algorithms have been developed. This paper reviews recent developments of deep learning techniques in the field of image-based weed detection. The review begins with an introduction to the fundamentals of deep learning related to weed detection. Next, recent progresses on deep weed detection are reviewed with the discussion of the research materials including public weed datasets. Finally, the challenges of developing practically deployable weed detection methods are summarized, together with the discussions of the opportunities for future research.We hope that this review will provide a timely survey of the field and attract more researchers to address this inter-disciplinary research problem.