A REVIEW OF UTILIZING DEEP LEARNING AND IMAGING FOR PLANT GROWTH ASSESSMENT

Abstract

This study evaluate advancements in the application of deep learning and multi-wavelength
imaging techniques for monitoring and phenotyping plant growth. With the rapid development of
plant breeding technology, effectively integrating high-throughput phenotyping platforms, utilizing
conventional imaging to tomographic imaging, which represent a significant step forward in
researching complex traits related to plant growth and adaptability. Although deep learning methods
have demonstrated breakthrough capabilities of image classification in various fields, their application
in plant growth monitoring presents challenges such as the need for extensive data annotation and the
ability to process spatial and temporal information simultaneously. This article emphasizes the
necessity of developing new softwares and techniques to improve data interpretability and achieves
results that align with plant physiological models. Progress in both deep learning and image techniques
areas promises to provide more detailed insights into plant phenotypes, accelerate analysis, and
enhance our understanding of plant development in diverse environments. This research not only
reviews a new state-of-the-art deep learning and image techniques but also provides comments on
the need for technology development and proposes future research directions.