Deep Learning for Image-Based Plant Growth Monitoring: A Review

Abstract

Deep learning (DL) approaches have received extensive attention in plant growth monitoring due to their ground-breaking performance in image classification; however, the approaches have yet to be fully explored. This review article, therefore, aims to provide a comprehensive overview of the work and the DL developments accomplished over the years. This work includes a brief introduction on plant growth monitoring and the image-based techniques used for phenotyping. The bottleneck in image analysis is discussed and the need of DL methods in plant growth monitoring is highlighted. A number of research works focused on DL based plant growth monitoring-related applications published since 2017 have been identified and included in this work for review. The results show that the advancement in DL approaches has driven plant growth monitoring towards more complicated schemes, from simple growth stages identification towards temporal growth information extraction. The challenges, such as resource-demanding data annotation, data-hungriness for training, and extraction of both spatial and temporal features simultaneously for accurate plant growth prediction, however, remain unsolved.

Deep learning-empowered crop breeding: intelligent, efficient and promising

“…Consequently, it has tremendous potential in predicting plant growth, estimating yield, detecting maturity, and perceiving biotic/abiotic stresses. However, deep learning algorithms require a large amount of labeled data ( Cordts et al., 2016 ), and data acquisition comes at a high cost, especially when identifying numerous categories ( Tong et al., 2022 ) or subtle differences between categories ( Taghavi Namin et al., 2018 ). Furthermore, collecting phenotype data faces additional obstacles of severe occlusion and various lighting conditions ( Scharr et al., 2016 ), which increase the time required for obtaining the necessary annotations.…”

Current State, Data Requirements and Generative AI Solution for Learning-based Computer Vision in Horticulture

“…Similarly, sequential models, using long short-term memory (LSTM) or recurrent neural networks (RNNs), capture temporal dependencies in plant growth using time sequence data. A recent survey [7] has identified a total of 23 published works proposing deep learning models for plant monitoring applications, from which 15 were for horticulture crops, with most of the works published between 2017 to 2021 using either CNN/R-CNN with spatial data or convolutional LSTM/RNN with spatiotemporal data. All of these works were primarily focused on either of the three objective tasks: (i) plant growth classification, (ii) instance segmentation and object detection for growth stage identification, and (iii) regression for plant growth analysis.…”

Recognition of Ginger Seed Growth Stages Using a Two-Stage Deep Learning Approach

“…Despite the use of DL methods in agriculture-related applications has been reported to be on the rise since 2016, the work to exploit the potential of DL methods in analyzing plant growth traits is still lacking in general compared to other applications in agriculture such as species classification, stress detection, and yield estimation [7]. Specifically, growth monitoring studies constitute only 14.08% of the 71 studies as reported by Yang and Xu [8].…”