Skip to content

Home

Jump to bottom
Sharon Fitzpatrick edited this page Jun 12, 2023 · 15 revisions

official_coastseg_logo

Useful Links

What is CoastSeg?

Coastseg stands for Coastal Segmentation, it is an interactive jupyter notebook for downloading satellite imagery with CoastSat and applying image segmentation models to satellite imagery. CoastSeg provides an interactive interface for drawing Regions of Interest (ROIs) on a map, downloading satellite imagery, loading geojson files, extracting shorelines from satellite imagery, and more.

  • A mapping extension for CoastSat using Segmentation Zoo models.
  • An interactive interface to download satellite imagery using CoastSat from Google Earth Engine
  • An interactive interface for extracting shorelines from satellite imagery
  • An interactive interface to apply segmentation models to satellite imagery

gif of map with rectangles on it

CoastSeg is a web browser tool for framing satellite image data extraction problems

  • It is an interactive mapping extension for CoastSat using Segmentation Zoo models
  • CoastSeg facilitates satellite-derived shoreline mapping using CoastSat functionality, in a web-browser for greater interactivity and data visualization options, providing a web mapping interface to prepare data downloads, view data, analyze segmentation outputs, and subsequent analyses
  • CoastSeg also facilitates the application of Segmentation Zoo models for customized mapping of coastal landscapes, using deep learning based image segmentation on coastal satellite imagery

Highlights

  1. CoastSeg will interface with, and extend the functionality of, CoastSat and CoastSat.Slope by
    • facilitating deep learning based image segmentation on coastal satellite imagery, and
    • providing a web mapping interface to prepare data downloads, view data, analyze segmentation outputs, and subsequent analyses

The hope is to provide a standardized workflow that does not involve external processes such as a GIS or file browser, so analysis can be carried out on a cloud computer, among other potential advantages.

  1. CoastSeg will facilitate the application of Segmentation Zoo models for deep-learning-based image segmentation on geospatial imagery with a web mapping interface. Our hope is that the use of deep learning models would facilitate custom mapping of coastal features beyond the current scope of CoastSat.

  2. CoastSeg will eventually facilitate custom image segmentation and mapping, by helping implement models that you have trained using Segmentation Gym and Doodler models

Clone this wiki locally