🌍

Hardi Andon 2024 Intro and Problem Statements

Jul 17, 2024

Hardi Andon 2024 Intro and Problem Statements

General Instructions

  • All focal points' presentations should be limited to 10 minutes.
  • Present problem statement, datasets provided, expected outcomes, suggested tools, and approaches.

Problem Statement 10: Lunar Crater Search

Presenters

  • Sur and Aay from Saad

Topic

  • Image-based search for lunar crater on the global mosaic of the lunar surface.

Details

  • Dataset: Lunar Global Mosaic (LROC) at 100m spatial resolution, freely available.
  • Task: Provide a lunar crater image and search the global mosaic to locate it with the crater's center latitude and longitude.
  • Tools: Python with PDS reader for interpreting Chandrayaan-2 Orbiter's Terrain Mapping Camera images.
  • Challenges: Handling spatial resolution differences (10m vs. 100m).

Problem Statement 11: Synthetic Image Generation Using DEM

Presenters

  • Sur and Canan

Topic

  • Generate synthetic lunar surface images using Digital Elevation Model (DEM).

Details

  • Dataset: DEM from Chandrayaan-2 is available. Additional info such as sun azimuth, elevation, and camera viewing parameters will be provided.
  • Task: Generate and visualize synthetic lunar surface images based on the DEM data and additional parameters.
  • Tools: Python package for DEM reading, PDS reader for data access.

Problem Statement 2: Solar Energy Estimation from Rooftops

Presenter

  • Sidat

Topic

  • Develop a web portal to estimate the annual solar energy generation potential of a rooftop.

Details

  • Dataset: Satellite images and solar radiation data will be provided; optionally use publicly available data for training models.
  • Task: Use ML/DL techniques to find building footprints and analyze potential solar energy generation based on satellite images and radiation data.
  • Tools: Python, Machine Learning Frameworks, UG (possibly PostGIS).
  • Expected Solution: Should estimate solar potential, offer a user-friendly selection interface.

Problem Statement 4: Voice-Enabled Geospatial Mapping

Presenter

  • Amit Bodani and Arpit Agarwal

Topic

  • Develop a voice-enabled user interface for geospatial map applications.

Details

  • Dataset: Use open GIS platforms and WMS services from NASA and Buoy.
  • Task: Recognize voice commands to manipulate and query geospatial map-based applications.
  • Tools: Python for back-end, JavaScript for UI, GIS libraries like Leaflet and Open Layers.
  • Challenges: Convert user spoken commands efficiently into actions, provide accurate and usable maps.

Problem Statement 5: Predicting Precipitation using Radar Data

Presenter

  • Abhishek

Topic

  • Develop an algorithm to nowcast precipitation systems using C-band weather radar observations.

Details

  • Dataset: Reflectivity scans from Doppler weather radar (NetCDF format).
  • Task: Identify, track, and predict precipitation systems and their movement in short-term forecasts.
  • Tools: MATLAB/Python, basic radar processing knowledge, Optical flow algorithms.

Problem Statement 6: Feature Extraction from High-Res Remote Sensing Data

Presenter

  • Santoshi and Kumar

Topic

  • Extract features (e.g., high tension towers, windmills) from high-resolution remote sensing data using AIML.

Details

  • Dataset: High-resolution satellite data like TIFF or JPEG, PDS standard data.
  • Task: Model feature extraction, segmentation, and build an interactive application for visualization and validation.
  • Tools: Python-based AIML frameworks, QGIS for labeling, Streamlit for UI.

Problem Statement 12: Context-Aware Spatial Data Retrieval

Presenter

  • Manindra

Topic

  • Design NLP/LLM system for spatial data queries and retrieval.

Details

  • Dataset: Annotated geospatial data, textual metadata from papers, articles, etc.
  • Task: NLP model for understanding spatial context in queries, retrieval of relevant geospatial data.
  • Tools: NLP Frameworks (SpaCy), GIS data libs, Machine learning frameworks.

Problem Statement 1: Digital Twin for Urban Area Traffic Simulation

Presenter

  • N Nant

Topic

  • Develop a digital twin for simulating and predicting urban traffic flow and congestion.

Details

  • Dataset: Open Street Map for detailed road networks.
  • Task: Simulate traffic, Predict impacts, assess scenarios, and visualize traffic data in a user interface.
  • Tools: Python, Sumo for simulation, PostGIS for data handling.

Problem Statement 3: Automatic Detection of Craters and Boulders

Presenter

  • Aditya and Suel

Topic

  • Develop automatic detection system for lunar boulders and craters from high-res orbital imagery.

Details

  • Dataset: OHRC images from Chandrayaan-2 at 0.25m resolution.
  • Task: Use imaging techniques to identify and map boulders and craters effectively.
  • Tools: Imaging and AI/ML tools.

Full transcript