DOTA航拍图像数据集

DescriptionDOTA is a large-scale dataset for object detection in aerial images. It can be used to develop and evaluate object detectors in aerial images. The images are collected from different sensors and platforms. Each image is of the size in the range from800 × 800to20,000 × 20,000pixels and contains objects exhibiting a wide variety of scales, orientations, and shapes. The instances in DOTA images are annotated by experts in aerial image interpretation by arbitrary (8 d.o.f.) quadrilateral. We will continue to update DOTA, to grow in size and scope to reflect evolving real-world conditions. Now it has three versions:DOTA-v1.0contains 15 common categories, 2,806 images and 188, 282 instances. The proportions of the training set, validation set, and testing set in DOTA-v1.0 are 1/2, 1/6, and 1/3, respectively.DOTA-v1.5uses the same images as DOTA-v1.0, but the extremely small instances (less than 10 pixels) are also annotated. Moreover, a new category, ”container crane” is added. It contains 403,318 instances in total. The number of images and dataset splits are the same as DOTA-v1.0. This version was released for theDOAI Challenge 2019on Object Detection in Aerial Images in conjunction with IEEE CVPR 2019.DOTA-v2.0collects more Google Earth, GF-2 Satellite, and aerial images. There are 18 common categories, 11,268 images and 1,793,658 instances in DOTA-v2.0. Compared to DOTA-v1.5, it further adds the new categories of ”airport” and ”helipad”. The 11,268 images of DOTA are split into training, validation, test-dev, and test-challenge sets. To avoid the problem of overfitting, the proportion of training and validation set is smaller than the test set. Furthermore, we have two test sets, namely test-dev and test-challenge. Training contains 1,830 images and 268,627 instances. Validation contains 593 images and 81,048 instances. We released the images and ground truths for training and validation sets. Test-dev contains 2,792 images and 353,346 instances. We released the images but not the ground truths. Test-challenge contains 6,053 images and 1,090,637 instances. The images and ground truths of test-challenge will be available only during the challenging.CitationIf you make use of the DOTA dataset, please cite our following paper:@misc{ding2021object,title={Object Detection in Aerial Images: A Large-Scale Benchmark and Challenges},author={Jian Ding and Nan Xue and Gui-Song Xia and Xiang Bai and Wen Yang and Micheal Ying Yang and Serge Belongie and Jiebo Luo and Mihai Datcu and Marcello Pelillo and Liangpei Zhang},year={2021},eprint={2102.12219},archivePrefix={arXiv},primaryClass={cs.CV}}@InProceedings{Xia_2018_CVPR,author = {Xia, Gui-Song and Bai, Xiang and Ding, Jian and Zhu, Zhen and Belongie, Serge and Luo, Jiebo and Datcu, Mihai and Pelillo, Marcello and Zhang, Liangpei},title = {DOTA: A Large-Scale Dataset for Object Detection in Aerial Images},booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},month = {June},year = {2018}}@InProceedings{Ding_2019_CVPR,author = {Jian Ding, Nan Xue, Yang Long, Gui-Song Xia, Qikai Lu},title = {Learning RoI Transformer for Detecting Oriented Objects in Aerial Images},booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},month = {June},year = {2019}}

DOTA是一款面向航空图像目标检测的大规模基准数据集，可用于开发与评估航空图像目标检测器。该数据集的图像采集自多种传感器与平台。每张图像的分辨率介于800×800至20000×20000像素之间，图像内的目标具备丰富的尺度、朝向与形态变化。DOTA数据集中的目标实例由航空图像解译专家采用任意8自由度（8 d.o.f.）四边形进行标注。团队将持续更新DOTA数据集，扩大其规模与覆盖范围，以贴合不断变化的真实场景需求。

目前该数据集共有三个版本：DOTA-v1.0包含15个通用类别，共计2806张图像与188282个目标实例。其训练集、验证集与测试集的划分比例分别为1/2、1/6与1/3。DOTA-v1.5沿用了DOTA-v1.0的全部图像，但新增了对极小目标实例（面积小于10像素）的标注，同时新增了"集装箱起重机"（container crane）这一类别，总目标实例数达403318个。该数据集的图像数量与划分方式与DOTA-v1.0完全一致，此版本是为配合2019年IEEE国际计算机视觉与模式识别会议（CVPR 2019）举办的2019年航空图像目标检测DOAI挑战赛而发布的。

DOTA-v2.0新增了Google Earth、高分二号（GF-2）卫星以及航空影像数据，共计包含18个通用类别、11268张图像与1793658个目标实例。相较于DOTA-v1.5，该版本进一步新增了"机场"与"直升机坪"两个类别。DOTA-v2.0的11268张图像被划分为训练集、验证集、测试开发集（test-dev）与测试挑战集（test-challenge）。为避免过拟合问题，训练集与验证集的占比低于测试集。该数据集共设有两类测试集：测试开发集与测试挑战集。其中训练集包含1830张图像与268627个目标实例，验证集包含593张图像与81048个目标实例，团队已公开训练集与验证集的图像及标注真值（ground truths）。测试开发集包含2792张图像与353346个目标实例，团队仅公开了该子集的图像，未提供标注真值。测试挑战集包含6053张图像与1090637个目标实例，其图像与标注真值仅在挑战赛期间对外开放。

引用须知：若您在研究中使用DOTA数据集，请引用以下论文：
@misc{ding2021object,title={Object Detection in Aerial Images: A Large-Scale Benchmark and Challenges},author={Jian Ding and Nan Xue and Gui-Song Xia and Xiang Bai and Wen Yang and Micheal Ying Yang and Serge Belongie and Jiebo Luo and Mihai Datcu and Marcello Pelillo and Liangpei Zhang},year={2021},eprint={2102.12219},archivePrefix={arXiv},primaryClass={cs.CV}}
@InProceedings{Xia_2018_CVPR,author = {Xia, Gui-Song and Bai, Xiang and Ding, Jian and Zhu, Zhen and Belongie, Serge and Luo, Jiebo and Datcu, Mihai and Pelillo, Marcello and Zhang, Liangpei},title = {DOTA: A Large-Scale Dataset for Object Detection in Aerial Images},booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},month = {June},year = {2018}}
@InProceedings{Ding_2019_CVPR,author = {Jian Ding, Nan Xue, Yang Long, Gui-Song Xia, Qikai Lu},title = {Learning RoI Transformer for Detecting Oriented Objects in Aerial Images},booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},month = {June},year = {2019}}

搜集汇总

背景与挑战

背景概述

DOTA航拍图像数据集是一个用于航拍图像目标检测的大规模数据集，包含从不同传感器和平台收集的图像，尺寸范围从800×800到20,000×像素，涵盖多种物体尺度和方向。数据集有三个版本：DOTA-v1.0有15个类别和2,806张图像，DOTA-v1.5增加了小实例标注和“container crane”类别，DOTA-v2.0扩展至18个类别和11,268张图像，并包含训练、验证和测试集划分，总计文件大小为156.33GB，适用于开发和评估目标检测算法。