Tutorials for the Low Altitude Disaster Imagery (LADI) dataset. This tutorial was originally forked from a Penn State Learning Factory capstone project.
We encourage the use of the GitHub Issues but when email is required, please contact the administrators at ladi-dataset-admin@mit.edu. As the public safety and computer vision communities adopt the dataset, a separate mailing list for development may be created.
This section specifies the run order and requirements for the initial setup the repository. Other repositories in this organization may be reliant upon this setup being completed.
Immediately after cloning this repository, create a persistent system environment variable titled LADI_DIR_TUTORIAL with a value of the full path to this repository root directory.
On unix there are many ways to do this, here is an example using /etc/profile.d. Create a new file ladi-env.sh using sudo vi /etc/profile.d/ladi-env.sh and add the command to set the variable:
export LADI_DIR_TUTORIAL=PATH TO /ladi-tutorialYou can confirm LADI_DIR_TUTORIAL was set in unix by inspecting the output of env.
This is a set of boilerplate scripts describing the normalized script pattern that GitHub uses in its projects. The GitHub Scripts To Rule Them All was used as a template. Refer to the script directory README for more details.
You will need to run these scripts in this order to download package dependencies and download all of the necessary data to get you started.
A set of tutorials focused on installing AWS tools and configuring AWS environment to download LADI dataset and load dataset in Python locally and remotely. There is also a short tutorial on how to clean and validate data.
A set of tutorials that are Jupyter Python 3.X notebooks that demonstrate on how to perform geospatial analysis by enhancing the LADI metadata with third party GIS information. One tutorial identifies the number of images taken within an administrative boundary (e.g. USA states) and assigns each state a color based on the number of images taken. The other tutorial filters images based on an specific annotation and performs various geospatial measurements on this subset.
These tutorials focus on how to training and testing a classifier model using Convolutional Neural Network (CNN) from scratch or using pre-trained ResNet and AlexNet.
This documentation is about loading LADI dataset in PyTorch framework including examples of writing custom Dataset, Transforms and Dataloader.
Please use this DOI number reference when citing the software:
