chessl/lakes

Introduction

This repo is part of a collaborative project aimed at classifying all 150,000 lakes in the continental US by the degree of human modification such as the presence of dams or water-control structures.

Project Structure

• cnn - Tensorflow code
  • data_load.py - Data Loader
  • tensorflow.py - CNN based on pure Tensorflow
  • tflearn.py - CNN based on TFlearn
• gee - Google Earth Engine code for feature extraction
  • exports - Feature exports
  • exports_iws - Feature exports for iws
• map_sampling - Map partitioning
  • main.py
  • samples.py
• ml - Machine Learning based on GEE features
  • models.ipynb