Course Objectives

• To study the spectral python tool for processing Hyperspectral images.
• To study the concept of Hyperspectral remote sensing.
• To know the basics, importance, and methods of Spectral remote sensing.
• To study machine learning technology for processing hyperspectral images.

Learning Outcomes

• To understand the structure of spectral data including band associations, shape and size of a hyperspectral images
• To develop software skills in programs used for map production in the modern cartographic workflow.
• To develop skills for deploy machine learning technology for processing multispectral and hyperspectral images.

Course Syllabus

Module 1

Practice:
1. Downloading multispectral satellite data from USGS/Bhuban etc
2. Installation of Python (Integrated Development and Learning Environment, Colab Notebook, Python 3.5 or above versions)

Module 2

Practice:
1. Reading multispectral data in python interface.
2. Reading hyperspectral data in pyhton interface.

Module 3

Practice:
1. installing dependencies library and visualization tools in Python such as Rasterio, shapely, GDIL, Spy, Matplotlib, Fiona, etc.
2. Check the shape and size of the image.

Module 4

Practice:
1. Importing Training dataset l.e Ground truth image of one band.
2. Display selected/filtred bands of Hyperspectral image.
3. Superimpose Ground truth image on Hyperspectral image.

Module 5

Practice:
1. Unsupervised classification of Hyperspectral image.

Module 6

Practice:
1. Supervised Classification of Hyperspectral image.

Module 7

Practice:
1. PCA for Dimensionality reduction of Hyperspecral image.
2. Minimum noise fraction hyperspectral data filtering.

Session Plan

Session 1

Data download from USGS and Bhuban, Hyperspectral data download from Weebly

Click here to watch the tutorial Video

Session 2

Python installation in window 10

Click here to watch the tutorial video

Session 3

Spectral Python Basics

Click here to watch the tutorial video

Session 4

Spectral Spectral Algorithms

Click here for more information

Projects

1. Training data set for supervised classification of Hyperspectral data.
2. Obtain spectral signatures for each of the bands of a Hyperspectral data.
3. PCA for dimensionality reduction
4. Unsupervised classification using the Gaussian classifier.

Case Studies

Applied Hyperspectral Imaging Fundamentals and Case Studies

Hyperspectral Image Case Studies