We are excited to announce the opening of our Landslide4Sense competition on April 1, 2022. Landslide4Sense focuses on innovative algorithms for landslide detection using multi-sensor satellite data.
Landslides are common in many parts of the world, but have become more frequent and damaging in recent years due to climate change and human activities. Often triggered by earthquakes or heavy rain falls, they cause adverse effects over large areas. Rapid and reliable landslide detection is critical for immediate response and rescue operations. Satellite images serve as the main data source for landslide detection.
Landslide4Sense provides a benchmark dataset of globally distributed multi-sensor satellite images. The dataset contains ~ 5000 image patches; each image patch is a composite of 14 bands, including multi-spectral data and slope and elevation information. The competition dataset is available for download.
The competition invites researchers in machine learning, computer vision, and remote sensing to develop innovative algorithms for automatic landslide detection. During phase 1 of the competition, participants receive the training and validation data and can rank their solutions in the competition leaderboard. During phase 2, participants receive the test data and are given a limited time to submit their solutions.
Landslide4Sense offers competitive prizes. The winners will be selected from the top three solutions in the competition leaderboard. Moreover, special prizes will be awarded for creative and innovative solutions.
The results of the Landslide4Sense competition will be presented during a special session of the 2nd workshop on Complex Data Challenges in Earth Observation (CDCEO) 2022. This year, CDCEO is part of the 31st International Joint Conference on Artificial Intelligence and the 25th European Conference on Artificial Intelligence (IJCAI-ECAI) running on July 23-29, 2022 in Vienna, Austria.
Are we supposed to submit the results in the similar H5 format that we get for the training and validation data? If so, we also submit the results of each image from the validation set individually, right?
Please post this question directly on the Landslide4sense competition forum: https://www.iarai.ac.at/landslide4sense/forums/
The organisers will be happy to help you!