Semantic Segmentation of Satellite Images

Capturing visible colors and a variety of other measurements from hundreds of miles away, satellite images have the potential to teach us a lot about the globe.

In this livestream, we’ll explain what makes satellite imagery so special and how you can start working with it. Then we will show how you can use AI to classify each pixel in those images, enabling us to gain a deeper understanding of what is shown and extracting information that may not be easy to see with the naked eye.  

To access the data shown today or get some hands-on practice with semantic segmentation, register for the Kelp Wanted challenge:  https://bit.ly/KelpWanted

Check out the following blog to access the code shown in this livestream: https://bit.ly/48uqMqN

Learn more about the tools and workflows shown in this video:
- What is Hyperspectral Imaging? https://bit.ly/Hyperspectral_Imaging
- The Hyperspectral Viewer App: https://bit.ly/HyperspectralViewer
- What is Semantic Segmentation? https://bit.ly/SemanticSegmentation
- Deep Network Designer App: https://bit.ly/DeepNetworkDesigner

About the presenters:
Grace Woolson graduated from Hamilton College with a degree in computer science in 2021, then joined MathWorks a few months later. Currently, she creates content that helps students learn how to use MATLAB for data science.

Abhishek Shankar graduated from Wayne State University with a master's in Robotics in 2023. Currently, he helps develop educational content for students to utilize MATLAB and Simulink for Robotics

Chapters:
0:18 Introduction
1:45 Understanding Satellite Images
4:45 Using the Hyperspectral Viewer App  
10:39 Exploring the non-spectral data
14:35 Importing Datasets
15:54 Splitting your data
16:42 Processing your data
23:45: Design and Train a Neural Network
31:11 Evaluating Semantic Segmentation Models
35:32 Create a Submission for Kelp Wanted
37:11 Closing

• #science_technology
• #matlab
• #simulink
• #mathworks
• #data_visualization
• #online_course
• #edx