FAIR1M
收藏资源简介:
FAIR1M数据集是由中国科学院航空信息研究所创建的,用于高分辨率遥感图像中的细粒度目标识别。该数据集包含超过15,000张图像和100万个实例,图像分辨率为0.3m至0.8m,覆盖多个国家和地区。FAIR1M数据集通过定向边界框对5个类别和37个子类别的目标进行标注,具有大规模、丰富的细粒度类别信息、地理信息和高质量图像等特点。该数据集旨在推动遥感图像中细粒度目标检测和视觉分类的研究,解决实际应用中的挑战。
The FAIR1M dataset was developed by the Aviation Data Information Research Institute of the Chinese Academy of Sciences for fine-grained object recognition in high-resolution remote sensing images. This dataset contains over 15,000 images and 1 million annotated instances, with image resolutions ranging from 0.3m to 0.8m, and covers multiple countries and regions. FAIR1M annotates objects across 5 main categories and 37 sub-categories using oriented bounding boxes, and is characterized by large scale, rich fine-grained category information, geospatial information, and high-quality imagery. This dataset aims to promote research on fine-grained object detection and visual classification in remote sensing images, and address challenges in practical applications.

- 1FAIR1M: A Benchmark Dataset for Fine-grained Object Recognition in High-Resolution Remote Sensing Imagery中国科学院航空信息研究所 · 2021年



