Abstract:Very-High Resolution (VHR) remote sensing imagery is increasingly accessible, but often lacks annotations for effective machine learning applications. Recent foundation models like GroundingDINO and Segment Anything (SAM) provide opportunities to automatically generate annotations. This study introduces FMARS (Foundation Model Annotations in Remote Sensing), a methodology leveraging VHR imagery and foundation models for fast and robust annotation. We focus on disaster management and provide a large-scale dataset with labels obtained from pre-event imagery over 19 disaster events, derived from the Maxar Open Data initiative. We train segmentation models on the generated labels, using Unsupervised Domain Adaptation (UDA) techniques to increase transferability to real-world scenarios. Our results demonstrate the effectiveness of leveraging foundation models to automatically annotate remote sensing data at scale, enabling robust downstream models for critical applications. Code and dataset are available at \url{<a class="link-external link-https" href="https://github.com/links-ads/igarss-fmars" rel="external noopener nofollow">this https URL</a>}.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: **Very - High - Resolution (VHR) remote sensing images lack high - quality annotations, which limits their effective use in supervised machine - learning applications**. Specifically, although VHR remote - sensing images are increasingly accessible, these images usually lack sufficient annotation information, especially in critical application scenarios such as disaster management. To solve this problem, the authors propose **FMARS (Foundation Model Annotations in Remote Sensing)**, a method for automatically generating large - scale annotations using foundation models. By combining VHR images with foundation models (such as GroundingDINO and Segment Anything), FMARS can quickly and robustly generate annotation data, thus supporting model training for downstream tasks (such as semantic segmentation and instance segmentation). ### Main problems and solutions 1. **Problems**: - Although VHR remote - sensing images are easily accessible, they lack high - quality annotations. - Existing annotation methods are inefficient and costly and are difficult to apply on a large scale. 2. **Solutions**: - **Introducing foundation models**: Utilize foundation models such as GroundingDINO and Segment Anything, which can provide powerful image classification, object detection, and segmentation capabilities without additional fine - tuning. - **Automating the annotation process**: Design an automated workflow to generate annotations from pre - event images, including three types of objects: buildings, roads, and high - vegetation. - **Constructing a large - scale dataset**: Based on the images of 19 disaster events provided by the Maxar Open Data program, construct a large - scale dataset containing more than 25 million annotations. - **Applying UDA technology**: In order to improve the quality of the generated annotations and the robustness of the model, unsupervised domain adaptation (UDA) technologies such as DAFormer and Masked Image Consistency (MIC) are adopted. ### Experimental verification To verify the effectiveness of FMARS, the authors trained the state - of - the - art segmentation models on the generated annotations and compared them with the manually annotated data. The experimental results show that even in a fully automated annotation process, the application of UDA technology significantly improves the performance of the model, especially when dealing with the challenging high - vegetation category. ### Conclusion FMARS provides an effective method for automatically generating high - quality annotations using foundation models, thereby solving the problem of insufficient annotation of VHR remote - sensing images. This method not only improves the annotation efficiency but also enables the training of accurate downstream segmentation models without a large amount of manual annotation.

FMARS: Annotating Remote Sensing Images for Disaster Management using Foundation Models

Enabling Quick, Accurate Crowdsourced Annotation for Elevation-Aware Flood Extent Mapping

Foundation Models for Remote Sensing and Earth Observation: A Survey

EMR-HRNet: A Multi-Scale Feature Fusion Network for Landslide Segmentation from Remote Sensing Images

A Modular Processing Chain for Automated Flood Monitoring from Multi-Spectral Satellite Data

SpaceNet 6: Multi-Sensor All Weather Mapping Dataset

Generative ConvNet Foundation Model With Sparse Modeling and Low-Frequency Reconstruction for Remote Sensing Image Interpretation

Towards a Knowledge guided Multimodal Foundation Model for Spatio-Temporal Remote Sensing Applications

A Machine learning approach for Post-Disaster data curation

High-Resolution Flood Monitoring Based on Advanced Statistical Modeling of Sentinel-1 Multi-Temporal Stacks

Improving Emergency Response during Hurricane Season using Computer Vision

Bridging Remote Sensors with Multisensor Geospatial Foundation Models

New Digital Field of Drawing and Survey for the Automatic Identification of Debris Accumulation in Flooded Areas

RescueADI: Adaptive Disaster Interpretation in Remote Sensing Images with Autonomous Agents

Locate Anything on Earth: Advancing Open-Vocabulary Object Detection for Remote Sensing Community

BlessemFlood21: Advancing Flood Analysis with a High-Resolution Georeferenced Dataset for Humanitarian Aid Support

Toward Foundation Models for Earth Monitoring: Generalizable Deep Learning Models for Natural Hazard Segmentation

A global multimodal flood event dataset with heterogeneous text and multi-source remote sensing images

AI Foundation Models in Remote Sensing: A Survey

Multi-Stage Synergistic Aggregation Network for Remote Sensing Visual Grounding