Deep Learning Application for Earth Observation

Course Curriculum

Chapter 1: Introduction

Lecture 1: Introduction

Lecture 2: Convolution neural network

Lecture 3: Resources to learn deep learning

Lecture 4: Notebooks and codes

Chapter 2: Python Basic

Lecture 1: Outline

Lecture 2: Python for absolutely beginners

Chapter 3: Deep learning environment setup

Lecture 1: Overview of this section

Lecture 2: Local environment vs Google Colab

Lecture 3: Anaconda and jupyter notebook installation

Lecture 4: Tensorflow installation

Lecture 5: Tensorflow in jupyter notebook

Lecture 6: Installation of matplotlib, gdal and rasterio

Lecture 7: Tensorflow in google colab

Chapter 4: Deep learning dataset preparation using ArcGIS Pro

Lecture 1: Required dataset for this lecture

Lecture 2: Create area of interest in ArcGIS pro

Lecture 3: Required code for upcoming section

Lecture 4: Download sentinel2 MSI imagery from GEE

Lecture 5: Create the output imagery (mask imagery)

Lecture 6: Create image and mask tiles using arcgis pro

Chapter 5: Open source solution for data preparation (geotile)

Lecture 1: Overview of this section

Lecture 2: About the input dataset

Lecture 3: Introduction to geotile package

Lecture 4: Correction on next lecture

Lecture 5: Image and Mask tiles using GeoTile

Chapter 6: Image classification

Lecture 1: Overview of this lecture

Lecture 2: Introduction to image classification

Lecture 3: About dataset

Lecture 4: Image classification deep learning model

Chapter 7: Deep learning object detection

Lecture 1: Overview of this section

Lecture 2: Object detection introduction

Lecture 3: About dataset

Lecture 4: Object detection using YOLOv4

Chapter 8: Image segmentation (Binary class)

Lecture 1: Overview of this section

Lecture 2: About dataset

Lecture 3: Building detection

Chapter 9: Image segmentation (Multi-class)

Lecture 1: Overview of this section

Lecture 2: Land use Land cover mapping

Lecture 3: Convert image to numpy file

Lecture 4: Convert numpy file to image

Chapter 10: Landslide detection

Lecture 1: Overview of this section

Lecture 2: Landslide inventory detection using CNN

Lecture 3: Attention U-Net introduction

Lecture 4: More information about attention unet

Lecture 5: Dataset using in below lecture

Lecture 6: Landslide detection using SAR dataset

Chapter 11: Time Series Analysis with LSTM

Lecture 1: Intro

Lecture 2: Get Data

Lecture 3: Data preprocessing

Lecture 4: LSTM model training and prediction

Chapter 12: Flood mapping

Lecture 1: Intro

Lecture 2: Data Preprocessing and Augmentation

Lecture 3: Attention unet and its implementation in tensorflow

Lecture 4: Training and evaluating the flood mapping model

Lecture 5: Hyperparameters fine tuning

Chapter 13: End to End Deep Learning and Google Earth Engine

Lecture 1: Overview of this section

Lecture 2: Objective of this section

Lecture 3: Download necessary satellite imagery

Lecture 4: Prepare image and mask file from area of interest

Lecture 5: Generate image and mask tiles

Lecture 6: Data normalization and save numpy array

Lecture 7: Test imagery and mask tiles and visualization of tiles

Lecture 8: Training UNet convolutional neural network

Lecture 9: Improvements on model (data augmentation, early stopping etc)

Lecture 10: Accuracy assessment

Lecture 11: Prediction of image and merge tiles as a single image

Lecture 12: Bonus lecture