Deep learning for plant stress detection: a comprehensive review of technologies, challenges, and future directions

Authors: PAUL, NIJHUM - North Dakota State University; GC, SUNIL - North Dakota State University; Horvath, David; SUN, XIN - North Dakota State University

Submitted to: Computers and Electronics in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 11/28/2024
Publication Date: 12/13/2024
Citation: Paul, N., Gc, S., Horvath, D.P., Sun, X. 2024. Deep learning for plant stress detection: A comprehensive review of technologies, challenges, and future directions. Computers and Electronics in Agriculture. https://doi.org/10.1016/j.compag.2024.109734.
DOI: https://doi.org/10.1016/j.compag.2024.109734

Interpretive Summary: Deep learning (DL)-based systems have emerged as powerful tools for the detection and management of plant stress, offering high accuracy and efficiency in analyzing imagery data. This review paper aims to present a thorough overview of the state-of-the-art in plant stress detection systems based on deep learning. For this purpose, a systematic literature review was conducted to identify relevant articles for highlighting the technologies and approaches currently employed in the development of a deep learning-based plant stress detection system, specifically the advancement of imagebased data collection systems, image preprocessing techniques and deep learning algorithms and their applications in plant stress classification, disease detection, and segmentation tasks. Additionally, this review emphasizes the challenges and future directions in collecting and preprocessing image data, model development, and deployment in real-world agricultural settings. Some of the key findings from this review paper are: (i) Training data collected from different growth settings or environmental conditions is important to increase the generalizability of the deep learning model; (ii) Most plant stress detection models have been trained on Red Green Blue (RGB) images; (iii) Overlapping stress symptoms can confuse the DL model; overlapping stress symptoms can be identified as a separate label such as “others” to solve this problem; (iv) Self-supervised learning (SSL) and Few-shot learning (FSL)-based methods may be better than transfer learning (TL)-based models for classifying plant stress when the number of labeled training images are scarce; (v) Data augmentation can increase both the quantity and variation of training data; (vi) Custom designed convolutional neural networks (CNN) architectures for a specific stress and plant type can have better performance than the state-of-the-art CNN architectures in terms of efficiency, overfitting, and accuracy; (vii) Handling multimodal inputs (e.g., image, temperature, humidity) allows the model to leverage information from diverse sources, which can improve prediction accuracy; (viii) The multi-task learning CNN structure reuses most of the network architecture while performing multiple tasks (e.g., estimate stress type and severity)simultaneously, which makes the learning much faster; (ix). There is a lack of research using unmanned aerial vehicle (UAV) or unmanned ground vehicle (UGV) imagery; and (x) Plant stress detection apps should have offline accessibility because remote field areas may not have internet access. This review can help to shape the development of novel approaches, strategies, and applications that address the present gaps and challenges in plant stress detection.

Technical Abstract: This review used a literature search to identify 37 unique publications in the use of artificial intelligence (AI) for agricultural purposes, specifically for stress responses. As such, it provides an overview of image acquisition and preprocessing technologies for plant stress detection, and a discussion of potential challenges and future directions in deep learning-based plant stress detection systems. This review will provide a comprehensive starting point for other researchers interested in using AI technologies to identify archetypes of plant stress and quantify damage caused by stress in plants.