Puzhao Zhang: Remote sensing for wildfire detection

Welcome to this week’s Learning Machines seminar. Title: Multi-Source Satellite Remote Sensing for Large-Scale Wildfire Burned Area Mapping Using Deep Learning Speaker: Puzhao Zhang, KTH Abstract: Wildfires are growing in intensity, frequency and extent around the globe, satellite remote sensing provides a cost-effective way to monitor wildfires and assess their environmental impacts over vast geographical areas, from local to global scale. Wildfire mapping and monitoring has long been hampered by the lack of data in high temporal and spatial resolution. A new era was started by the public release of 50 years’ Landsat data, and then boosted by the launch of Sentinel-1/2 satellites, making satellite remote sensing play an increasingly important role in understanding environmental dynamics happening on our planet. Optical multispectral observations have long been used for mapping wildfires and assessing their impacts. Despite the satisfactory results produced by optical observations, difficulties may arise due to the presence of cloud cover, smoke, and haze, which could significantly reduce the observation frequency in the visible/infrared domain. SAR can penetrate through clouds and smoke, being capable of imaging day and night at all weather, therefore has great potential to complement optical observations. This talk will introduce our recent work on exploiting multi-source satellite remote sensing data for large-scale wildfire burned area mapping with deep learning. About the speaker: Dr. Puzhao Zhang, researcher from the Division of Geoinformatics, KTH. He received his PhD degree in Geoinformatics from KTH in 2021. His postdoctoral position was funded by the Digital Futures Postdoctoral fellowship. He is co-editing a special issue “Multi-source Data with Remote Sensing Techniques” for the journal Remote Sensing. His research interests include satellite remote sensing big data analytics, geospatial machine learning, and environmental change monitoring. More information and coming seminars: https://ri.se/lm-sem – The Learning Machines Team