Skip to content
Showcase/DOTA Aerial Image Object Detection
advancedimage

DOTA Aerial Image Object Detection

Oriented bounding box detection in aerial images (Xia et al., CVPR 2018). Detect 15 object categories with arbitrary orientations including planes, ships, vehicles, and sports facilities.

Labels:outdoornatureurbanpeopleanimal+−

Configuration Fileconfig.yaml

This Potato config reproduces the annotation task. Save it as config.yaml and run potato start config.yaml to try it.

yaml
# DOTA Aerial Image Object Detection Configuration
# Based on Xia et al., CVPR 2018

annotation_task_name: "DOTA Aerial Object Detection"
task_dir: "."

data_files:
  - "sample-data.json"

item_properties:
  id_key: "id"
  text_key: "image_url"
  context_key: "context"

user_config:
  allow_all_users: true

annotation_schemes:
  - annotation_type: "multiselect"
    name: "object_classes"
    description: "Select all object classes visible"
    labels:
      - name: "plane"
        tooltip: "Airplanes"
      - name: "ship"
        tooltip: "Ships and vessels"
      - name: "storage_tank"
        tooltip: "Storage tanks"
      - name: "baseball_diamond"
        tooltip: "Baseball diamonds"
      - name: "tennis_court"
        tooltip: "Tennis courts"
      - name: "basketball_court"
        tooltip: "Basketball courts"
      - name: "ground_track_field"
        tooltip: "Running tracks"
      - name: "harbor"
        tooltip: "Harbors"
      - name: "bridge"
        tooltip: "Bridges"
      - name: "large_vehicle"
        tooltip: "Large vehicles (trucks, buses)"
      - name: "small_vehicle"
        tooltip: "Small vehicles (cars)"
      - name: "helicopter"
        tooltip: "Helicopters"
      - name: "roundabout"
        tooltip: "Roundabouts"
      - name: "soccer_field"
        tooltip: "Soccer fields"
      - name: "swimming_pool"
        tooltip: "Swimming pools"

  - annotation_type: "radio"
    name: "difficulty"
    description: "Annotation difficulty"
    labels:
      - name: "easy"
        tooltip: "Clear, large objects"
      - name: "difficult"
        tooltip: "Small, occluded, or crowded"

  - annotation_type: "text"
    name: "oriented_bbox"
    description: "Oriented bounding box: x1,y1,x2,y2,x3,y3,x4,y4,class,difficulty"

interface_config:
  item_display_format: "<img src='{{text}}' style='max-width:100%; max-height:500px;'/><br/><small>{{context}}</small>"

output_annotation_format: "json"
output_annotation_dir: "annotations"

Sample Datasample-data.json

json
[
  {
    "id": "dota_001",
    "image_url": "https://upload.wikimedia.org/wikipedia/commons/thumb/1/10/Empire_State_Building_%28aerial_view%29.jpg/800px-Empire_State_Building_%28aerial_view%29.jpg",
    "context": "Aerial image. Detect objects with oriented bounding boxes. Objects may appear at any angle."
  },
  {
    "id": "dota_002",
    "image_url": "https://upload.wikimedia.org/wikipedia/commons/thumb/1/1e/San_Francisco_from_the_Marin_Headlands_in_March_2019.jpg/1200px-San_Francisco_from_the_Marin_Headlands_in_March_2019.jpg",
    "context": "Aerial view. Mark all DOTA categories: planes, ships, vehicles, sports facilities, etc."
  }
]

// ... and 1 more items

Get This Design

View on GitHub

Clone or download from the repository

Quick start:

git clone https://github.com/davidjurgens/potato-showcase.git
cd potato-showcase/image/aerial/dota-aerial
potato start config.yaml

Dataset & paper

Xia et al., CVPR 2018

Official dataset ↗Read the paper ↗

Citation (BibTeX)

bibtex
@inproceedings{xia2018dota,
    title = "{DOTA}: A Large-scale Dataset for Object Detection in Aerial Images",
    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",
    booktitle = "Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition",
    pages = "3974--3983",
    year = "2018",
    url = "https://arxiv.org/abs/1711.10398"
}

Details

Annotation Types

multiselectradiotext

Domain

Remote SensingAerial Imagery

Use Cases

Object DetectionOriented DetectionAerial Analysis

Tags

dotaaerialoriented-bboxremote-sensingcvpr2018

Found an issue or want to improve this design?

Open an Issue

Related Designs

FLAIR: French Land Cover from Aerospace Imagery

Land use and land cover classification from high-resolution aerial imagery. Annotators classify the primary land use category of aerial image patches and identify any secondary land uses present. Based on the FLAIR dataset from the French National Institute of Geographic and Forest Information (IGN).

multiselectradio

xView Satellite Object Detection

Large-scale overhead imagery object detection (Lam et al., arXiv 2018). Detect 60 object classes including vehicles, buildings, and infrastructure from satellite images.

multiselectradio

CUB-200-2011: Fine-Grained Bird Species Dataset

CUB-200-2011 is a fine-grained bird image dataset of 11,788 photos across 200 species, each labeled with parts, attributes, and a bounding box (Wah et al., 2011). This Potato config reproduces the labeling task.

multiselectradio