Towards Geography-Specific Benchmarking in AI/ML for geospatial datasets: A Focus on LiDAR Datasets

Towards Geography-Specific Benchmarking in AI/ML for geospatial datasets: A Focus on LiDAR Datasets

ABSTRACT:
Geospatial Artificial Intelligence (GeoAI) integrates spatial data with AI/ML techniques to support decision-making in areas such as urban planning, environmental monitoring, and disaster management. Among these spatial data sources, LiDAR (Light Detection and Ranging) offers precise 3D point cloud data critical for extracting objects (e.g., buildings, vegetation, terrain). However, most AI/ML models trained on shape-based geometric features from LiDAR data suffer from a lack of geographic diversity, as they are predominantly developed using datasets from urban regions in Europe or North America. This geographic bias limits the models’ ability to generalize across different environments, resulting in reduced accuracy and increased bias when applied to underrepresented regions—ultimately compromising their reliability in critical application. This talk will highlight the need for region specific benchmark datasets which will enable the development of geographically robust AI/ML models, improving the generalizability across diverse terrains.

13th June 2025

@ 3:30 PM IST

Meeting link:
https://bit.ly/4iWdBUl

Meeting ID: 496 378 381 343 1
Passcode: UY9Zk2co

Speaker

Dr. A.M.Ramiya

Dr. A. M. Ramiya

Associate Professor
Indian Institute of Space Science and Technology Department of Space
Thiruvananthapuram, Kerala

Dr A M Ramiya is an Associate Professor in Remote Sensing and Image Processing at the Indian Institute of Space Science and Technology, Department of Space, Thiruvananthapuram.

She has received her PhD in Remote Sensing from IIST, Thiruvananthapuram, and M.S. in Remote Sensing and Spatial Analysis, University of Southampton, UK and B E in Geoinformatics from College of Engineering, Guindy.

She has more than 16 years of experience in handling and processing geospatial images from multispectral/ hyperspectral/ LiDAR sensors captured from various platforms such as satellite, airborne, drone/UAV, ground-based and mobile platforms.

Her research interests include developing algorithms and methodologies using machine learning techniques for processing very high-resolution remote sensing images for natural and man-made resource management.

She is the recipient of the prestigious Commonwealth award from the Government of UK for pursuing Masters program in the United Kingdom. Recently she has been awarded the Kerala State Young Scientist award 2021 from the Government of Kerala for her contribution in Geospatial Technology. Currently she acts as the Vice Chair of IEEE GRSS Kerala Chapter.