Скачать или смотреть How LANDFIRE Uses Image-Based Modeling to Map Vegetation & Update Fuels | LANDFIRE Office Hour

Join this LANDFIRE Office Hour Daryn Dockter, LANDFIRE Technical Lead, TSSC, EROS explains the many steps that must occur to update and create LANDFIRE's annually updated product suite. Daryn discusses how machine learning and modeling are used to create the three primary vegetation products including Existing Vegetation Type (EVT), Existing Vegetation Cover (EVC), and Existing Vegetation Height (EVH). He explains the process involving satellite image compositing, field plot data compilation, model generation and iteration, and the incorporation of land cover data in the final vegetation map products.

0:00 Intro
2:06 Introducing terms, concepts, AI, ML, Neural Nets, Deep Learning...
3:52 Machine learning - supervised learning vs. unsupervised learning
6:30 How we use ML in LANDFIRE to model vegetation (explained)
10:20 Building the model (explained)
11:52 Labeling each pixel to dominant lifeform (explained)
15:20 Creating LF vegetation cover and height modeling with ML
19:50 Creating LF vegetation type with ML
25:04 LF version timeline map
28:00 How to handle mixed pixels when modeling dominant lifeforms?
28:55 Differences in published datasets in 2016 and current datasets?
30:00 What's the future of updates, especially Existing Vegetation Type?
31:28 Garbage in....garbage out? How to ensure high-quality data in (and) outputs
32:42 Why is Harmonized Landsat Sentinel main input for the model?
34:48 Explored other remote sensing techniques for training data aside from Lidar?
36:20 Dealing with the "difficult" types on EVT (discussed)
39:55 The importance of plot classification - where plot data are robust vs. lacking
43:26 Predictions made by season for wildfires?
44:58 Enhancing training and validation data through citizen science?



Want more short-form LANDFIRE Info? Join the email list http://eepurl.com/baJ_BH
Subscribe: YouTube channel:    / landfirevideo  
Twitter: @nature_LANDFIRE   / nature_landfire  
Find LANDFIRE data and products https://landfire.gov/
Still have a question? Shoot us an email [email protected] (we're here to help!)

#LANDFIRE #LANDFIRECOMMUNITY #LANDFIREHELP #LANDFIREHOWTO