article · Frontiers in Plant Science
Accurate detection of plant diseases under real-world field conditions is vital for effective crop management. This research investigates the use of the SegFormer deep learning model to perform precise semantic segmentation of multiple strawberry diseases from natural images. Researchers evaluated three Mix Transformer encoders, specifically MiT-B0, MiT-B3, and MiT-B5, across seven common strawberry afflictions, including leaf spots, rots, blights, and powdery mildew on leaves and fruit. A dataset of 2,450 raw images was expanded through augmentation to 4,574 images, assisted by the Segment Anything Model for annotation. Testing demonstrated that MiT-B3 and MiT-B5 consistently surpassed MiT-B0 in segmentation accuracy. While MiT-B3 showed steady adaptation, MiT-B5 achieved the highest overall segmentation precision, offering a framework that could be extended to other agricultural crops.
Identifying crop diseases directly in natural field settings is challenging due to variable lighting and complex backgrounds. By showing how transformer-based vision models accurately segment specific plant lesions, this work helps agricultural technologists select the right computer vision architectures. Improving automated disease recognition supports earlier diagnosis and more targeted crop treatments, which can reduce yield losses in fruit farming.
This work is relevant to developers of smart agriculture platforms, automated crop monitoring systems, and digital agronomy tools. The research demonstrates an applied and tested computer vision pipeline for identifying specific strawberry diseases from natural photos. While demonstrated effectively on experimental image sets, further engineering would be required to integrate these models into field-deployable drone, tractor, or smartphone scouting applications for commercial growers.
AI-generated from the published abstract. Always read the original work before citing.
Introduction: Precise semantic segmentation of microbial alterations is paramount for their evaluation and treatment. This study focuses on harnessing the SegFormer segmentation model for precise semantic segmentation of strawberry diseases, aiming to improve disease detection accuracy under natural acquisition conditions. Methods: Three distinct Mix Transformer encoders - MiT-B0, MiT-B3, and MiT-B5 - were thoroughly analyzed to enhance disease detection, targeting diseases such as Angular leaf spot, Anthracnose rot, Blossom blight, Gray mold, Leaf spot, Powdery mildew on fruit, and Powdery mildew on leaves. The dataset consisted of 2,450 raw images, expanded to 4,574 augmented images. The Segment Anything Model integrated into the Roboflow annotation tool facilitated efficient annotation and dataset preparation. Results: The results reveal that MiT-B0 demonstrates balanced but slightly overfitting behavior, MiT-B3 adapts rapidly with consistent training and validation performance, and MiT-B5 offers efficient learning with occasional fluctuations, providing robust performance. MiT-B3 and MiT-B5 consistently outperformed MiT-B0 across disease types, with MiT-B5 achieving the most precise segmentation in general. Discussion: The findings provide key insights for researchers to select the most suitable encoder for disease detection applications, propelling the field forward for further investigation. The success in strawberry disease analysis suggests potential for extending this approach to other crops and diseases, paving the way for future research and interdisciplinary collaboration.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.3389/fpls.2024.1352935
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
