Coast Train: Massive Library of Labeled Coastal Images to Train Machine Learning for Coastal Hazards and Resources
Active
By Community for Data Integration (CDI)
April 2, 2021
Scientists who study coastal ecosystems and hazards such as hurricanes, flooding, and cliff failure collect lots of photographs of coastal environments from airplanes and drones. A large area can be surveyed at high resolution and low cost. Additionally, satellites such as Landsat have provided imagery of the Nation’s coastlines every few days for decades. Scientist’s ability to understand coastal hazards would be greatly improved if this wealth of imagery could be ‘mined’ automatically by computers. We want to automate the process of identifying and labelling each region of the image from a set of categories (e.g. bare land, water, woody vegetation, herbaceous vegetation). We need to train a computer to recognize the same things that a human can and that will require a very large number of labeled images. We’d like to work with STEP-UP students to do most of the labelling using software we write especially for the purpose.
Principal Investigator : Phillipe A Wernette
Co-Investigator : Sara L Zeigler, Nicholas M Enwright
Cooperator/Partner : Daniel D Buscombe, Evan Goldstein
Principal Investigator : Phillipe A Wernette
Co-Investigator : Sara L Zeigler, Nicholas M Enwright
Cooperator/Partner : Daniel D Buscombe, Evan Goldstein
- Source: USGS Sciencebase (id: 60677364d34edc0435c09d57)
Coast Train--Labeled imagery for training and evaluation of data-driven models for image segmentation
Coast Train is a library of images of coastal environments, annotations, and corresponding thematic label masks (or 'label images') collated for the purposes of training and evaluating machine learning (ML), deep learning, and other models for image segmentation. It includes image sets from both geospatial satellite, aerial, and UAV imagery and orthomosaics, as well as non-geospatial...
A 1.2 billion pixel human-labeled dataset for data-driven classification of coastal environments
The world’s coastlines are spatially highly variable, coupled-human-natural systems that comprise a nested hierarchy of component landforms, ecosystems, and human interventions, each interacting over a range of space and time scales. Understanding and predicting coastline dynamics necessitates frequent observation from imaging sensors on remote sensing platforms. Machine Learning models...
Authors
Daniel Buscombe, Phillipe Alan Wernette, Sharon Fitzpatrick, Jaycee Favela, Evan B. Goldstein, Nicholas Enwright
Scientists who study coastal ecosystems and hazards such as hurricanes, flooding, and cliff failure collect lots of photographs of coastal environments from airplanes and drones. A large area can be surveyed at high resolution and low cost. Additionally, satellites such as Landsat have provided imagery of the Nation’s coastlines every few days for decades. Scientist’s ability to understand coastal hazards would be greatly improved if this wealth of imagery could be ‘mined’ automatically by computers. We want to automate the process of identifying and labelling each region of the image from a set of categories (e.g. bare land, water, woody vegetation, herbaceous vegetation). We need to train a computer to recognize the same things that a human can and that will require a very large number of labeled images. We’d like to work with STEP-UP students to do most of the labelling using software we write especially for the purpose.
Principal Investigator : Phillipe A Wernette
Co-Investigator : Sara L Zeigler, Nicholas M Enwright
Cooperator/Partner : Daniel D Buscombe, Evan Goldstein
Principal Investigator : Phillipe A Wernette
Co-Investigator : Sara L Zeigler, Nicholas M Enwright
Cooperator/Partner : Daniel D Buscombe, Evan Goldstein
- Source: USGS Sciencebase (id: 60677364d34edc0435c09d57)
Coast Train--Labeled imagery for training and evaluation of data-driven models for image segmentation
Coast Train is a library of images of coastal environments, annotations, and corresponding thematic label masks (or 'label images') collated for the purposes of training and evaluating machine learning (ML), deep learning, and other models for image segmentation. It includes image sets from both geospatial satellite, aerial, and UAV imagery and orthomosaics, as well as non-geospatial...
A 1.2 billion pixel human-labeled dataset for data-driven classification of coastal environments
The world’s coastlines are spatially highly variable, coupled-human-natural systems that comprise a nested hierarchy of component landforms, ecosystems, and human interventions, each interacting over a range of space and time scales. Understanding and predicting coastline dynamics necessitates frequent observation from imaging sensors on remote sensing platforms. Machine Learning models...
Authors
Daniel Buscombe, Phillipe Alan Wernette, Sharon Fitzpatrick, Jaycee Favela, Evan B. Goldstein, Nicholas Enwright