arXiv reaDer
Performance evaluation of deep segmentation models for Contrails detection
Contrails, short for condensation trails, are line-shaped ice clouds produced by aircraft engine exhaust when they fly through cold and humid air. They generate a greenhouse effect by absorbing or directing back to Earth approximately 33% of emitted outgoing longwave radiation. They account for over half of the climate change resulting from aviation activities. Avoiding contrails and adjusting flight routes could be an inexpensive and effective way to reduce their impact. An accurate, automated, and reliable detection algorithm is required to develop and evaluate contrail avoidance strategies. Advancement in contrail detection has been severely limited due to several factors, primarily due to a lack of quality-labeled data. Recently, proposed a large human-labeled Landsat-8 contrails dataset. Each contrail is carefully labeled with various inputs in various scenes of Landsat-8 satellite imagery. In this work, we benchmark several popular segmentation models with combinations of different loss functions and encoder backbones. This work is the first to apply state-of-the-art segmentation techniques to detect contrails in low-orbit satellite imagery. Our work can also be used as an open benchmark for contrail segmentation and is publicly available.
updated: Sat Nov 04 2023 19:10:38 GMT+0000 (UTC)
published: Sun Nov 27 2022 15:07:39 GMT+0000 (UTC)
参考文献 (このサイトで利用可能なもの) / References (only if available on this site)
被参照文献 (このサイトで利用可能なものを新しい順に) / Citations (only if available on this site, in order of most recent)
Amazon.co.jpアソシエイト