BigEarthNet Remote Sensing Classification

Multi-label land cover classification from Sentinel-2 imagery (Sumbul et al., IGARSS 2019). Classify satellite patches into 43 land cover classes following the CORINE taxonomy.

📝

text annotation

Configuration Fileconfig.yaml

# BigEarthNet Remote Sensing Classification Configuration
# Based on Sumbul et al., IGARSS 2019

annotation_task_name: "BigEarthNet Land Cover Classification"

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: "artificial_surfaces"
    description: "Select visible artificial surfaces"
    labels:
      - name: "urban_fabric"
        tooltip: "Continuous and discontinuous urban fabric"
      - name: "industrial"
        tooltip: "Industrial or commercial units"
      - name: "road_rail"
        tooltip: "Road and rail networks"
      - name: "port"
        tooltip: "Port areas"
      - name: "airport"
        tooltip: "Airports"
      - name: "construction"
        tooltip: "Construction sites"
      - name: "green_urban"
        tooltip: "Green urban areas"
      - name: "sport_leisure"
        tooltip: "Sport and leisure facilities"

  - annotation_type: "multiselect"
    name: "agricultural"
    description: "Select agricultural land types"
    labels:
      - name: "arable_land"
        tooltip: "Arable land (crops)"
      - name: "permanent_crops"
        tooltip: "Vineyards, fruit trees, olive groves"
      - name: "pastures"
        tooltip: "Pastures"
      - name: "heterogeneous"
        tooltip: "Mixed agricultural areas"

  - annotation_type: "multiselect"
    name: "forest"
    description: "Select forest types"
    labels:
      - name: "broad_leaved"
        tooltip: "Broad-leaved forest"
      - name: "coniferous"
        tooltip: "Coniferous forest"
      - name: "mixed_forest"
        tooltip: "Mixed forest"

  - annotation_type: "multiselect"
    name: "natural"
    description: "Select natural vegetation and open spaces"
    labels:
      - name: "natural_grassland"
        tooltip: "Natural grasslands"
      - name: "moors_heathland"
        tooltip: "Moors and heathland"
      - name: "sclerophyllous"
        tooltip: "Sclerophyllous vegetation"
      - name: "transitional_woodland"
        tooltip: "Transitional woodland-shrub"
      - name: "bare_rock"
        tooltip: "Bare rocks"
      - name: "sparsely_vegetated"
        tooltip: "Sparsely vegetated areas"

  - annotation_type: "multiselect"
    name: "water"
    description: "Select water bodies"
    labels:
      - name: "inland_waters"
        tooltip: "Inland waters"
      - name: "marine_waters"
        tooltip: "Marine waters"
      - name: "coastal_lagoons"
        tooltip: "Coastal lagoons"
      - name: "estuaries"
        tooltip: "Estuaries"

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

[
  {
    "id": "ben_001",
    "image_url": "https://upload.wikimedia.org/wikipedia/commons/thumb/e/e7/Everest_North_Face_toward_Base_Camp_Tibet_Luca_Galuzzi_2006.jpg/1200px-Everest_North_Face_toward_Base_Camp_Tibet_Luca_Galuzzi_2006.jpg",
    "context": "Sentinel-2 satellite patch (120x120 pixels). Classify all visible land cover types according to CORINE Land Cover taxonomy."
  },
  {
    "id": "ben_002",
    "image_url": "https://upload.wikimedia.org/wikipedia/commons/thumb/1/1a/24701-nature-702704.jpg/1200px-24701-nature-702704.jpg",
    "context": "Satellite image patch. Select all applicable land cover classes."
  }
]

// ... 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/bigearth-net
potato start config.yaml

Details

Annotation Types

multiselect

Domain

Remote SensingEarth Observation

Use Cases

Land Cover ClassificationSatellite Image AnalysisEnvironmental Monitoring

Related Designs

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.

bboxmultiselect

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.