The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)

Using Deep Learning to Predict Wildfires with Feng Yan - #329

Dec 20, 2019
Feng Yan, an Assistant Professor at the University of Nevada, Reno, is at the forefront of using machine learning for wildfire prediction. He introduces ALERTWildfire, a network of cameras that capture real-time data to enhance monitoring efforts. The conversation dives into innovative camera deployments, the integration of satellite and ground-level data, and overcoming challenges in model training. Feng also discusses leveraging IaaS and FaaS for scalability and cost-effectiveness in tackling the growing threat of wildfires.
Ask episode
AI Snips
Chapters
Transcript
Episode notes
INSIGHT

Limitations of Satellite Imagery

  • Satellite imagery, while used for wildfire monitoring, has limitations.
  • Its low resolution makes it unsuitable for tracking rapidly spreading smoke.
ANECDOTE

Manual Wildfire Monitoring

  • Feng Yan's collaborators deployed cameras to monitor wildfires.
  • They manually checked for fires, lacking an automated alert system.
INSIGHT

Challenges in Smoke Detection

  • Wildfire smoke detection faces challenges due to varied shapes and environmental factors.
  • Differentiating smoke from clouds in mountainous areas is difficult, even for humans.
Get the Snipd Podcast app to discover more snips from this episode
Get the app