1 Introduction

1.1 Big Data

1.2 Geospatial Big Data

NASA was generating 12.1TB of data every single day (in 2017)

Satellites provide over 150 terabytes per day Source: ESA

Landsat Data

1.3 Cloud Computing

Can you estimate how many Landsat scenes are needed to cover this area?

What difficulties arise when the study area is large? Landsat 9, 01.7.2024 - 17.7.2024

1.4 Google Earth Engine

1.5 Getting started with GEE

The Google Earth Engine Interface

JavaScript Basic in GEE

1.6 Exploring GEE Data Catalog

Raster vs. Vector

gsp.humboldt.edu

Raster

Resolution: there are four types of resolution: radiometric, spatial, spectral, and temporal. “Spatial resolution relates to pixel size, temporal resolution to observation frequency, radiometric resolution to the number of unique values, and spectral resolution to binwidth in the electromagnetic spectrum.”

Mahood et a. (2023)

Landsat 8 vs. Sentinel 2

Vector

1.7 Import data to GEE

Import drone imagery and shapefile provided in “Data” session to your GEE account.

1.8 Export data from GEE

2 Data Management, Analysis, and Applications

2.1 Remote Sensing Indices

Remote indices (spectral indices) are used to enhance particular land surface features or properties, e.g. vegetation, soil, water, urban, fires…

Water Indices - NDWI

2.2 Cloud Masking

2.3 Reducer

2.4 Image Composition

Sentinel 2 (Median Composition, Jun-Jul 2020)

3 Land Cover Classification in GEE

3.1 Select/Identify Area of Interest

3.2 Prepare Input Data

Phan et al. 2020

3.3 Collect Samples

Different sampling approaches and their major advantages and drawbacks.

A simple, fast, and accurate method

Phan et al. 2022

3.4 ML Classifier

Phan et al. 2018

3.5 Accuracy Assessment

4 Time series