A new basic model dubbed RingMo has been developed to improve the accuracy of remote sensing image interpretation, according to the Aerospace Information Research Institute (AIR) of the Chinese Academy of Sciences (CAS).
Remote sensing images have been successfully applied in many fields, such as classification and change detection, and deep learning approaches have contributed to the rapid development of remote sensing (RS) image interpretation. The most widely used training paradigm is to use pre-trained ImageNet models to process RS data for specified tasks.
However, there are problems such as the domain gap between natural and RS scenes, and the poor generalization ability of RS models. It makes sense to develop a base model with a general representation of RS characteristics. Since a large amount of unlabeled data is available, the self-supervised method has more importance for development than the fully supervised method in remote sensing.
The study aims to propose a basic model framework of remote sensing, which can leverage the benefits of generative self-supervised learning for RS images. RingMo offers a large-scale dataset built by collecting two million RS images from satellite and aerial platforms, covering multiple scenes and objects around the world. Moreover, the RS base model training method is designed for dense and small objects in complex RS scenes.
RingMo is the first generative base model for cross-modal remote sensing data. It is state-of-the-art across eight datasets spread across four downstream tasks, demonstrating the effectiveness of the proposed framework. In the future, the model can be applied to 3D reconstruction, residential construction, transportation, water conservancy, environmental protection and other fields.
Research report:RingMo: a basic remote sensing model with masked image modeling
We need your help. The SpaceDaily news network continues to grow, but revenue has never been harder to sustain.
With the rise of ad blockers and Facebook, our traditional sources of revenue through quality network advertising continue to decline. And unlike so many other news sites, we don’t have a paywall – with those boring usernames and passwords.
Our news coverage takes time and effort to publish 365 days a year.
If you find our news sites informative and useful, consider becoming a regular supporter or, for now, making a one-time contribution.
SpaceDaily Monthly Supporter
$5+ billed monthly
$5 charged once
credit card or paypal
Albedo raises $48 million to capture highest resolution satellite imagery
Austin Texas (SPX) Sep 08, 2022
Albedo, a company developing low-flying satellites that will provide ultra-high resolution imagery, announced a $48 million Series A funding round co-led by Breakthrough Energy Ventures and Shield Capital, raising funding company total to $58 million in less than two years. since its creation. Participation in the round included new investors Republic Capital, Giant Step Capital and C16 Ventures, as well as existing investors Initialized Capital, Joe Montana’s Liquid 2, Kevin Mahaffey and others not disc…read more