Workshop: Unlocking the Power of Satellite Data with Deep Learning

*** Lunch should be taken care by the attendee ***

*** The registration of this workshop is closed and confirmation email sent to all registered participants***

* This is a beginner level workshop and participants are expected to bring their laptop

In this workshop, we dive deep into the world of satellite data and its vast potential. From environmental monitoring to urban planning and disaster response, satellite imagery has become a critical tool in various industries. This hands-on workshop will focus on leveraging deep learning models to analyze satellite data effectively.

1. Introduction to Satellite Data (20 mins)

• Overview of satellite imagery (optical, SAR, multispectral, etc.).
• Applications of satellite data across industries.

2. Basics of Deep Learning for Satellite Data (30 mins)

• Introduction to convolutional neural networks (CNNs) for image analysis.
• Popular frameworks and tools: TensorFlow, PyTorch.
• Pre-trained models and transfer learning.

3. Hands-on Session: Building a Satellite Data Classifier (60 mins)

• Dataset preparation (e.g., from Sentinel-2 or Landsat).
• Annotating and preprocessing satellite images.
• Training a CNN model to identify land cover types.

4. Advanced Techniques (45 mins)

• Semantic segmentation with U-Net for change detection.
• Object detection models (YOLO, Faster R-CNN) for infrastructure analysis.
• Time-series analysis using recurrent networks (LSTMs).

5. Use Case Discussion and Challenges (30 mins)

• Real-world projects leveraging satellite data and AI.
• Challenges in satellite data analysis (e.g., resolution, cloud cover).

6. Q&A and Closing Remarks (15 mins)

• Open floor for questions and discussions.
• Networking and resource sharing.

#### Who Should Attend:

• Data scientists and AI enthusiasts interested in remote sensing.
• Professionals in geospatial, environmental, and urban planning sectors.
• Students and researchers looking to explore AI applications in satellite data.