← Marketplace

Building Detector

Detects buildings for each pixel in the input image it returns the probability map of it to be a part of a building.

Description

The Building Detector identifies all pixels corresponding to buildings in satellite images. The block runs on Pléiades Display (recommended) or SPOT Display data blocks and outputs a mask with the probability of each pixel corresponding to a building. The solution uses deep learning techniques for computer vision and is constantly improved with subsequent updates. Thus far, it has been trained on tens of thousands of semantically annotated images.

Supported Workflows

This block can be used with:

  • Recommended: Pléiades Display (Streaming) -> Raster Tiling -> Building Detector
  • SPOT Display (Streaming) -> Raster Tiling -> Building Detector
  • Recommended: Pléiades Display (Download) -> DIMAP -> GeoTIFF Conversion -> Raster Tiling -> Building Detector
  • SPOT Display (Download) -> DIMAP -> GeoTIFF Conversion -> Raster Tiling -> Building Detector

Performance

The block uses Spacept’s custom computer vision model (deep neural network with an Augmented Feature Pyramid Architecture) trained on thousands of Pleiades images where Buildings have been semantically annotated. The model accuracy ranges from 80 to 90% depending on input block resolution and location. This is real-world accuracy following benchmarking against helicopter based LIDAR. Benchmarked against human annotators on 10,000+ Pleiades images the model showed an accuracy up to 90% depending upon the area and we are continously improving the model. The model was trained using labeled images taken from but not limited to parts of the United States, Mexico, France, Belgium, Portugal, India, Nepal, Indonesia, and Australia.

Input Parameters

This block currently can be used with Pléiades or SPOT Display. We recommend the use of the Raster Tiling block (512 x 512 tile size) upstream to speed up processing and reduce infrastructure costs for large areas. Future data block integrations will come in subsequent versions.

Output Format

GeoTIFF activation map image containing the probability for each pixel to be part of a building.

Use Cases

Use cases include Vegetation Management, Urban Planning & Construction, Fire Risk Estimation, Land Use and Management, Powerline inspection.

Video Tutorial

Spacept has created a 5-minute demonstration on running a job using the Building Detector Block, which is accessible in this YouTube tutorial video.

Example Input

{
   "oneatlas-pleiades-aoiclipped:1":{
      "ids":null,
      "bbox":[
         -1.039205,
         46.732618,
         -1.026172,
         46.743978
      ],
      "time":"2018-01-01T00:00:00+00:00/2020-12-31T23:59:59+00:00",
      "limit":1,
      "zoom_level":18,
      "time_series":null,
      "max_cloud_cover":100,
      "panchromatic_band":false
   },
   
    "tiling:1":{
      "nodata":null,
      "tile_width":512,
      "tile_height":512,
      "match_extents":false,
      "output_prefix":"",
      "augmentation_factor":1,
      "discard_empty_tiles":true
   },
   
    "building-detection:1":{
   }
}

Capabilities

To know more please check the block capabilities specifications.

Terms & Conditions

View the End User License Agreement conditions.

Ready to get started?

Get in touch or become a partner.

Contact salesBecome a partner