2026-10-14T21:05:21Z https://www.scidb.cn/oai

10.57760/sciencedb.32371 2026-04-13T10:13:10Z

2026-04-13 IAS-16k-D doi:10.57760/sciencedb.32371 en To support object-level detection of invasive alien animals and insects, we constructed the IAS-16k-D (Invasive Alien Species 16k-Detection) dataset for invasive alien species detection.Using LabelImg, we performed object-level bounding box annotation on 5,537 images covering invasive alien animals and insects, and uniformly converted all annotations to the YOLO format.The dataset was strictly split into training, validation, and test sets at a ratio of 70:15:15, supporting direct training of modern YOLO-series detectors (including YOLOv5, YOLOv8, YOLOv11, and YOLO26) without additional format conversion.IAS-16k-D covers five major taxonomic groups — Mollusca, Reptilia, Insecta, Fish, and Amphibia — including 18 representative invasive alien species, providing comprehensive coverage for research and real-time field monitoring of invasive alien animals and insects.The dataset is organized in a standard YOLO directory structure for out-of-the-box use, with the following components:train/: Training set containing images/ (training images) and labels/ (YOLO-formatted .txt annotation files), used for model training.val/: Validation set with the same structure as the training set, used for performance validation and hyperparameter tuning during training.test/: Test set used for independent and objective evaluation of final detection performance.classes.txt: Category index file listing the standard names of the 18 invasive alien species, fully compatible with the YOLO labeling format.source.json: Metadata file recording the original source, license type, and attribution information of each image to ensure compliant data usage. Invasive Alien Species; Invasive Alien Species Detection; IAS Guoxi Liu PUBLIC https://creativecommons.org/licenses/by-nc/4.0/ dataset Science Data Bank