COFGA
收藏资源简介:
COFGA数据集是由以色列国防部国防研究与发展局创建的,专注于从航空影像中进行细粒度物体分类。该数据集包含2,104张图像,这些图像是从航空成像系统中收集的,具有5-15厘米的地面采样距离(GSD),提供了比大多数公共航空影像数据集更高的空间分辨率。数据集中的14,256个标注物体被分类为2个主要类别、15个子类别、14个独特特征和8种感知颜色,总计37个不同的标签,使其非常适合细粒度分类任务。COFGA数据集的应用领域包括航空影像分析和利用,旨在解决细粒度物体分类的挑战。
The COFGA dataset was developed by the Defense Research and Development Directorate of the Israeli Ministry of Defense, focusing on fine-grained object classification from aerial imagery. This dataset comprises 2,104 images collected from aerial imaging systems, with a ground sampling distance (GSD) of 5–15 cm, providing higher spatial resolution than most public aerial image datasets. The 14,256 annotated objects within the dataset are classified into 2 main categories, 15 subcategories, 14 unique features, and 8 perceptual colors, amounting to 37 distinct labels in total, which renders it highly suitable for fine-grained classification tasks. The application domains of the COFGA dataset include aerial imagery analysis and utilization, with the goal of addressing the challenges in fine-grained object classification.

- 1cofga: A Dataset for Fine Grained Classification of Objects from Aerial Imagery以色列国防部国防研究与发展局 · 2021年



