Skip to content

Latest commit

 

History

History
34 lines (23 loc) · 1.29 KB

land-cover-classification.md

File metadata and controls

34 lines (23 loc) · 1.29 KB
Raw

Team project assignment - Land cover classification with Deep Learning using Sentinal-2 multispectral data

Research goal and expected output

  • The goal of this assignment is to use multi-spectral satellite data from Sentinal-2 to create a classification model to detect 9 or 10 land use classes (pasture, anual crop, residential etc.) and test it on Slovakia land data.

  • The advantage of using Sentinal-2 data is that it has a high revisit time (5 days) and is free available.

  • This model has pottential to be used in many tasks (to monitor forest logging etc.)

  • Expected output

  • trained classification model using Keras and Python
  • find a way to access Slovakia land data
  • environmnent to predict land use of Slovakia every 5 days

Milestones

  1. week - dataset exploration and net design
  2. week - dataset preparation and first working implementation of the CNN
  3. week - training net
  4. week - evaluating models
  5. week - preparing data for predicting (access Slovakia land data)
  6. week - generate label map of the area

Literature

Contact information