ORIGINAL RESEARCH article
Front. Plant Sci.
Sec. Sustainable and Intelligent Phytoprotection
Volume 16 - 2025 | doi: 10.3389/fpls.2025.1632698
This article is part of the Research TopicPlant Phenotyping for AgricultureView all 3 articles
A sorghum seed variety identification method based on image-hyperspectral fusion and an improved deep residual convolutional network
Provisionally accepted
- 1Changchun University of Science and Technology, Changchun, China
- 2Jilin Engineering Normal University, Changchun, Jilin, China
- 3Jilin Academy of Agricultural Sciences (CAAS), Changchun, Jilin Province, China
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Identifying sorghum seed varieties is crucial for ensuring seed quality, increasing planting efficiency, and advancing sustainable agricultural growth as a significant food and feed crop. This paper suggests a classification model based on spectrogram fusion and an improved deep residual convolutional network for quick and nondestructive classification of sorghum seeds of various types to achieve high-precision classification of sorghum seeds. A spectrogram fusion dataset comprising 12,800 seeds from eight distinct sorghum seed varieties was created for this study. An improved residual network model was then suggested, which included the addition of the deep separable convolution (DSC) and the attention-based mechanism (CBAM) to the ResNet network architecture. The findings demonstrate the high classification task performance of the CBAM-ResNet50-DSC model based on the spectrogram fusion dataset, with classification accuracy of 94.84%, specificity of 99.20%, recall of 94.39%, Precision of 94.52%, and F1-score of 0.9438. This study shows that the suggested network model can successfully classify sorghum seeds, offering a dependable and scientific way to quickly and non-destructively identify sorghum seeds.
Keywords: artificial intelligence, sorghum seed, Variety identification, Multi-modal fusion, ResNet models
Received: 21 May 2025; Accepted: 23 Jul 2025.
Copyright: © 2025 Yang, Chen, Song and Zhang. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Shaozhong Song, Jilin Engineering Normal University, Changchun, Jilin, China
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.