Page 686 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

also help generate reliable, labeled training datasets to further refine and guide excavation
strategies.

Data Collection and Field Validation:

Extended field validation will be essential to confirm and fine-tune the predictions generated
through remote sensing and machine learning. Planned field verification and ground-truthing
surveys will ensure alignment with archaeological evidence. In addition, supporting geophysical
methods—such as GPR and magnetic surveys—will be integrated to enhance subsurface
validation. Further data collection will also include acquiring updated satellite imagery to
support long-term site monitoring and improve the quality of training datasets.

Technology and Sensor Upgrades:

The project will benefit from expanded access to advanced remote sensing instruments,
including thermal sensors and GPR, to complement the existing SAR and multispectral datasets.
Integration of multitemporal and multisensor data sources will improve the detection of seasonal
and environmental variations, enhancing the overall robustness of the prediction models.

Scalability and Regional Expansion:

The methodology developed at the Saruq Al-Hadid site will be adapted for application in
similar arid environments across the Arabian Peninsula and globally. This scalability will help
increase the impact of the solution and expand its contribution to the understanding and
preservation of archaeological landscapes in desert and environmentally challenging settings.
Interdisciplinary and Institutional Collaboration:

Future work will continue to foster collaboration between remote sensing experts, archaeologists,
and machine learning specialists to align technological development with heritage conservation
priorities. Expanding partnerships with academic institutions and research centers in related
fields will strengthen the scientific basis of the work. Collaborations with technology providers
and data agencies can support access to up-to-date, high-quality imagery and remote sensing
resources.

3 Use Case Requirements

REQ-01: High-Resolution Imagery Acquisition
Must obtain high-resolution multispectral imagery (e.g., Worldview-3) to capture detailed
surface features.

REQ-02: SAR Data Integration
Must incorporate SAR data (e.g., ALOS-2/PALSAR-2) to enable subsurface feature detection
in arid environments.

REQ-03: Advanced Image Processing Capabilities
Must support advanced techniques such as PCA and spectral band transformations to enhance
and extract relevant features.

REQ-04: Geospatial Data Integration and Analysis
Must integrate geospatial datasets (e.g., digital elevation models, slope maps, hydrographic
networks) and perform spatial analysis using GIS tools.

650