LADI-Dataset/ladi-tutorial: September 2020

This repository is a collection of tutorials on how to access, analyze, and train a classifier using the LADI dataset.

Computer vision capabilities have rapidly been advancing and are expected to become an important component to incident and disaster response. However, the majority of computer vision capabilities are not meeting public safety’s needs, such as support for search and rescue, due to the lack of appropriate training data and requirements. For example in 2019, a leading computer vision benchmark has mislabeled a flooded region as a “toilet,” or a highway surrounded by flooding as a “runway.” In response, we’ve developed a dataset of images collected by the Civil Air Patrol of various disasters. The raw images were previously released into the public domain. Two key distinctions are the low altitude, oblique perspective of the imagery and disaster-related features, which are rarely featured in computer vision benchmarks and datasets. A subset of images were annotated using Amazon MTurk and high confidence results were achieved by consensus of qualified workers, who were evaluated on their ability to recognize objects via a qualification test. The dataset currently employs a hierarchical labeling scheme of a five coarse categorical and then more specific annotations for each category. The initial dataset focuses on the Atlantic Hurricane and spring flooding seasons since 2015. We also provide annotations produced from the commercial Google Cloud Vision service and open source Places365 benchmark.