DIKU Bits: Deep learning and remote sensing for ecosystem monitoring

Portrait of Christian Igel


Christian IgelProfessor in the Machine Learning section (ML)


Deep learning and remote sensing for ecosystem monitoring


Deep learning applied to remote sensing data allows for large-scale mapping of individual trees and forests, which enables more accurate estimation of captured carbon. This talk demonstrates tree crown segmentation in satellite images using fully convolutional neural networks (CNNs) and how allometric equations can then be applied to estimate biomass. Then it is shown how CNNs for 3D point clouds can estimate tree biomass directly based on LiDAR (laser imaging, detection, and ranging).

Which courses do you teach?

Machine Learning A (MLA)
Machine Learning B (MLB)
Online and Reinforcement Learning (OReL)

Which technology/research/projects/startup are you excited to see the evolution of?

I am just in the process of co-founding a startup in the domain of medical data analysis, and I am very excited about it.