2026-10-15T21:41:22Z https://data.mendeley.com/oai

oai:data.mendeley.com/65rzpcgbp4.3 2026-09-14T13:26:39Z

Rafi, Rafiullah Soybean Seed Image Dataset for Healthy and Damaged Seed Detection Mendeley Data This dataset contains original soybean seed images developed for computer vision and object detection research, with a focus on the detection of healthy and damaged soybean seeds. The dataset consists of images captured using an Infinix HOT 60i smartphone under controlled imaging conditions, with soybean seeds placed on a white paper background. The dataset contains two classes: healthy seeds and damaged seeds. The original images are provided in their original form, together with corresponding annotations for object detection. A 640 × 640 pixel version of the images is also provided as a model-ready format for deep learning applications. The dataset is organized into training, validation, and testing subsets. The original dataset contains 1,511 images, distributed as 1,209 training images, 151 validation images, and 151 testing images. The dataset was prepared and annotated using Roboflow. The annotations identify individual soybean seed instances according to their visual condition. The dataset is intended to support research in soybean seed inspection, agricultural computer vision, object detection, deep learning, and automated seed quality assessment. This dataset is associated with the research article “Automated Detection of Healthy and Damaged Soybean Seeds Using YOLO26.” The dataset is provided to support reproducibility and further research in automated soybean seed detection. Agricultural Science Computer Vision Deep Learning Image Reconstruction Dataset doi:10.17632/65rzpcgbp4.3 oai:data.mendeley.com/65rzpcgbp4.3 Creative Commons Attribution 4.0 International http://creativecommons.org/licenses/by/4.0 https://data.mendeley.com/datasets/65rzpcgbp4 2026-09-14T13:26:39Z