Locust Images Dataset

This a our own curated dataset of locust images with labels(as per YOLO format) that we used in our research paper to evaluate the performance of two state-of-the-art object dtecetion algorithams YOLOv5 and YOLOv8. So, this dataset was curated by gathering several locust images and these images were classified into two main categories on the basis of image background. 

Images with artificial and blurred background. These images are taken from roboflow website.

Images with natural backgroud with visible soil. These images are taken from the link https://lcas.lincoln.ac.uk/nextcloud/index.php/s/gGiGUxhYTmPV1vK

Then we use some augmentation techniques on this curated dataset and made the dataset more vast and versatile.

本数据集为我们自研的带标注蝗虫图像精选数据集，标注格式符合YOLO格式（YOLO format），曾用于我们的研究论文，以评估两款当前顶尖的目标检测算法（object detection algorithm）——YOLOv5与YOLOv8的性能。
本数据集通过收集多张蝗虫图像构建，并依据图像背景分为两大主要类别：
带有人工模糊背景的图像：此类图像采集自Roboflow官方网站。
带有可见土壤的自然背景图像：此类图像来源于以下公开链接：https://lcas.lincoln.ac.uk/nextcloud/index.php/s/gGiGUxhYTmPV1vK
随后，我们对该精选数据集应用了多种数据增强技术，进一步扩充了数据集规模并提升了其多样性与泛用性。