Abstract:Image segmentation is a crucial task in computer vision, which divides a digital image into multiple segments and objects. In agriculture, image segmentation is extensively used for crop and soil monitoring, predicting the best times to sow, fertilize, and harvest, estimating crop yield, and detecting plant diseases. However, image segmentation faces difficulties in agriculture, such as the challenges of disease staging recognition, labeling inconsistency, and changes in plant morphology with the environment. Consequently, we have conducted a comprehensive review of image segmentation techniques based on deep learning, exploring the development and prospects of image segmentation in agriculture. Deep learning-based image segmentation solutions widely used in agriculture are categorized into eight main groups: encoder-decoder structures, multi-scale and pyramid-based methods, dilated convolutional networks, visual attention models, generative adversarial networks, graph neural networks, instance segmentation networks, and transformer-based models. In addition, the applications of image segmentation methods in agriculture are presented, such as plant disease detection, weed identification, crop growth monitoring, crop yield estimation, and counting. Furthermore, a collection of publicly available plant image segmentation datasets has been reviewed, and the evaluation and comparison of performance for image segmentation algorithms have been conducted on benchmark datasets. Finally, there is a discussion of the challenges and future prospects of image segmentation in agriculture.

What problem does this paper attempt to address?

This paper attempts to solve several key problems in agricultural image segmentation, which are crucial for agricultural production and crop management. Specifically, the paper mainly focuses on the following aspects: 1. **Disease - stage identification**: In plant disease detection, diseases at different stages may show similar features in images, which makes it difficult to use image data for disease - stage identification. The paper explores how to improve the accuracy of disease - stage identification through deep - learning methods. 2. **Label inconsistency**: The inconsistent label standards in the training dataset may lead to inaccurate model training and evaluation results. The paper discusses how to reduce the impact of label inconsistency by improving the annotation standards and data augmentation techniques. 3. **Plant morphological changes**: Environmental factors can cause changes in the texture, size, and shape of plant leaves and roots, which pose challenges to the robustness and accuracy of image segmentation. The paper studies how to deal with these changes through multi - scale and attention mechanism methods. 4. **Comprehensive review and classification**: The paper conducts a comprehensive review of deep - learning - based image - segmentation techniques and divides them into eight main structures: encoder - decoder structure, multi - scale and pyramid methods, dilated convolutional networks, visual attention models, generative adversarial networks, graph neural networks, instance - segmentation networks, and Transformer - based models. The characteristics and application scenarios of each method are discussed in detail. 5. **Agricultural applications**: The paper introduces the specific applications of image - segmentation techniques in agriculture, including plant - disease detection, weed identification, crop - growth monitoring, crop - yield estimation and counting, etc. Through the analysis of these applications, the paper shows the potential and practical effects of image - segmentation techniques in the agricultural field. 6. **Public datasets and performance evaluation**: The paper also reviews several public plant - image - segmentation datasets and evaluates and compares the performance of various image - segmentation algorithms on the benchmark datasets, providing references for researchers. 7. **Future challenges and prospects**: Finally, the paper discusses the challenges faced by agricultural image segmentation and future research directions, including how to further improve the robustness and accuracy of models and how to better utilize large - scale data and advanced deep - learning techniques. In general, this paper aims to provide systematic guidance and references for researchers in the agricultural field by comprehensively reviewing and analyzing existing deep - learning image - segmentation techniques, and promote the development and application of agricultural image - processing techniques.

Deep learning implementation of image segmentation in agricultural applications: a comprehensive review

Deep learning-based instance segmentation architectures in agriculture: A review of the scopes and challenges

A Review of Deep Learning in Multiscale Agricultural Sensing

A review of deep learning used in the hyperspectral image analysis for agriculture

Applications of Deep Learning for Dense Scenes Analysis in Agriculture: A Review

Review of Crop Disease and Pest Detection Algorithms Based on Deep Learning

A review of remote sensing image segmentation by deep learning methods

A Research Review on Deep Learning Combined with Hyperspectral Imaging in Multiscale Agricultural Sensing

Sementing the Field of Rapeseed from 3D Laser Point Cloud Using Deep Learning

A systematic review of the use of Deep Learning in Satellite Imagery for Agriculture

Segmentation of Green Vegetation in the Field Using Deep Neural Networks*

Plant Disease Detection and Classification by Deep Learning—A Review

A systematic review of deep learning techniques for plant diseases

Farmland Segmentation from Remote Sensing Images Using Deep Learning Methods

Intelligent Agriculture: Deep Learning in UAV-based Remote Sensing Imagery for Crop Diseases and Pests Detection

Intelligent Image Semantic Segmentation: A Review Through Deep Learning Techniques for Remote Sensing Image Analysis

Deep Learning-based Agricultural Pest and Disease Recognition

A deep learning image segmentation model for agricultural irrigation system classification

Deep learning techniques for hyperspectral image analysis in agriculture: A review

Deep learning in agriculture: A survey

Deep learning in image segmentation for mineral production: A review