Image Analytics in Google Earth Engine

This repository contains Google Earth Engine JavaScript codes for the Image Analytics Course at COGS. The code is organized by assignment and includes codes for performing supervised machine learning classification, regression modeling, and change detection. You can also load the repository in the Earth Engine Code Editor: https://code.earthengine.google.com/?accept_repo=users/cindyhopp/gee-git

A1 - Supervised Machine Learning Classification

In this assignment, machine learning classifiers including CART, RF, and SVM were used to perform supervised land cover classifications on a Sentinel-2 image centred around Cochrane, Alberta, Canada. The code includes data preprocessing steps, model training, and accuracy assessment. The accuracy of the classification is assessed using confusion matrices, producer's and user's accuracy, and the overall accuracy of the model.

A2 - Regression Modeling

In this assignment, various types of regression models were used to predict tree cover based on multispectral imagery and analyze their performances. The Area of Interest (AOI) was set in Waterton Park, Alberta, where major forest fire took place in 2017. The code includes data preprocessing steps, model training, and accuracy assessment. The accuracy of the regression models is assessed using R-squared values and visual inspection of the predicted vs. observed tree cover.

A3 - Time Series Analysis and Change Detection

In this assignment, a change detection and time series analysis was completed in and around Beijing, the capital of China, during summer months between 2013 and 2021. The code includes data preprocessing steps, change detection algorithms, and visualization products. The time series visualization products include spectral plots of vegetation indices and maps showing the extent of change over time. A report with a more structured discussion of the project is also included in this repository.

Acknowledgments

This repository was developed as part of the Geospatial Data Analytics Program at Centre of Geographic Sciences, NSCC. Thank-you to Rob Hodder for their guidance and support in remote sensing, and the development team of the Cloud-Based Remote Sensing with Google Earth Engine: Fundamentals and Applications for a comprehensive guidebook.