Abstract
This talk will cover two key research areas being explored in my lab. In the field of forest research, we are developing a complete workflow, from conducting plot-level sample surveys using terrestrial laser scanning (TLS) to large-scale biomass mapping with aerial laser scanning. The discussion will focus on our work with TLS data, including leaf-filtering and volume estimation, and will demonstrate our progress in scaling plot-level data to broader regions for biomass mapping. Additionally, the talk will highlight a benchmarking study that assesses state-of-the-art deep learning models for forest segmentation and the development of a novel architecture tailored for this purpose. The second part of the talk will focus on our work in creating millions of scenarios and generating corresponding sensor data. This is designed to produce synthetic data for training and testing deep learning models deployed in autonomous systems, such as self-driving vehicles. Furthermore, the talk will briefly introduce the Geoinformatics Program at IIT Kanpur.
About the speaker
Bharat Lohani is a professor of Geomatics at IIT Kanpur, India and co-founder of Geokno India Pvt. Ltd. He received his PhD in remote sensing from the University of Reading, UK and a Master of Engineering in Photogrammetry and Remote Sensing at IIT Roorkee. Professor Lohani has been pioneering lidar research in India.
