Welcome to GEOL0069 AI for Earth Observation

Week 1

• Preparation
• Introduction to Intelligently Reinforced Image Segmentation (IRIS)

Week 2

• Introduction to AI/Machine Learning
• Core concepts
• Sea-ice and Lead Classification
• AI/Machine Learning Implementation

Week 3

• Data Fetching
• Roll-out on a Full Image
• Load in CNN
• Load in ViT
• Load in Random Forest
• Rollout on a Small Region
• Rollout on a full region

Week 4

• Colocating Sentinel-3 OLCI/SRAL and Sentinal-2 Optical Data
• Unsupervised Learning

Week 5

• Regression Techniques for Predictive Analysis
• Gaussian Processes bishop2006pattern
• Creating Training Data from Sentinel-2 and Sentinel-3 OLCI Data
• Application of Regression Techniques in Satellite Imagery Analysis

Week 6

• Imagery Alignment: Introduction and Simple Examples
• Images alignment on S3/S2
• Creating Training Data from Aligned and Misaligned Imagery
• Regression Application on Melt Pond Fractions (MPF)
• Gaussian Processes Regression Model

Week 7

• Introduction to Gaussian Processes
• Introduction to GPSat
• Using GPUs to accelerate training and inference
• GPSat modelling with local GP experts: A 1-D case study (Part 1)
• GPSat modelling with local GP experts: A 1-D case study (Part 2)
• 2D interpolation of ABC satellite data

Week 8

• GPSat Along-Track Interpolation
• Load data
• Select profile for interpolation
• Generate expert locations at 200 km
• Interpolate along-track using GPSat
• Interpolation of Sea Level Anomaly using GPSat
• Sea Level Anomaly Interpolation Using GPSat (GPOD)
• Finding Ocean Eddies using Satellite Altimetry: Part 1
• Finding Ocean Eddies using Satellite Altimetry: Part 2

Week 9

• Explainable AI
• Explainable AI Part 2
• Explainable AI in Polynomial Regression
• Neural Network