SatelliteFootprintDetection

This Project's aim is Footprint Detection of Buildings in High-Resolution Satellite Images by using instance segmentation.

Dataset Characteristics:

• This open-source dataset includes 24 images (one per month) covering ~100 unique geographies. 
• The dataset will comprise over 40,000 square kilometers of imagery and exhaustive polygon labels of building footprints in the imagery, totaling over 10 million individual annotations. 
• The Data – ~100 locations, spread out across the globe and contains:

Dataset :

• The Dataset is available for download on kaggle

• The Dataset is available for download on AWS as a Public Dataset:

  Training Data

  • aws s3 cp s3://spacenet-dataset/spacenet/SN7_buildings/tarballs/SN7_buildings_train.tar.gz .
  • aws s3 cp s3://spacenet-dataset/spacenet/SN7_buildings/tarballs/SN7_buildings_train_csvs.tar.gz .

  Testing Data

  • aws s3 cp s3://spacenet-dataset/spacenet/SN7_buildings/tarballs/SN7_buildings_test_public.tar.gz . Align center:

Models:

The Models trained are stored in GoogleDrive

BLOG POST:

The Whole Project is documented in this Blog Post

Results

results-video.mp4

Installation:

GIT:

.. code-block:: bash

  git clone https://github.com/PriyanK7n/SatFootprint
  cd SatFootprint
  pip install -r requirements.txt
  pip install -e .