SATELLITE-BASED ARCHAEOLOGICAL DISCOVERY

We see the history beneath the soil from space.

ARGEOS is a deep-learning powered discovery platform that detects undiscovered ancient settlements, mounds, and ruins by matching satellite imagery against a library of reference archaeological structures.

Apply for Early Access Explore the System
1.484 Reference structures
2.300+ Candidates detected
7,67 Area scanned (km²)
≥ 78% Classification accuracy
Map Overview
Triple Validation
Dashboard
// FEATURES

Beyond the capacity of a single researcher.

ARGEOS is a command console that compresses field reconnaissance from days into hours. The capabilities below are running in production right now.

Live Command Console

Monitor active scans, cells per second, ETA and last score on a single screen. The worker process self-heals on interruption.

Geographic Boundary + KML

Upload KML/KMZ from Google Earth or draw a polygon on the map — the scan never crosses these limits.

Reference Image Library

Build a reference set from known mounds, castles, monasteries and ruins. The system learns these motifs and uses them in search.

Grid Scan over Satellite Imagery

The region is divided into precise grid cells. Every cell is analyzed individually against high-resolution satellite imagery.

Interrupted Scan Recovery

Even after a power cut, cancellation, or system error, scans resume where they left off — no detected pins are ever lost.

3D Validation: Image + DEM

For each candidate: matched reference · high-res satellite detail · real SRTM/ASTER elevation model (DEM). Topographic check in one click.

Real-time Notifications

Receive an instant alert when a high-confidence candidate is detected. Every finding is logged with its pin number and score.

KML Export + Field Validation

Download candidate pins as KML compatible with Google Earth/QGIS. Your field team gets the coordinates directly.

// IN ACTION

Case Study: Anatolian Mid-Pliocene Mounds

In a 7.67 km² test area within the Konya–Karaman corridor, 1,199 grid cells were scanned. In approximately 5 hours, more than 10 high-confidence candidate structures were identified — the majority being mound candidates absent from the known literature.

Dashboard

Dashboard

Real-time system status, active scans and live metrics.

Map Overview

Map Overview

Grid scan over high-resolution satellite imagery.

Reference Library

Reference Library

Trained structure catalogs — categorized classification.

Region Definition

Region Definition

KML upload or polygon drawing on the map.

Interrupted Scan Recovery

Interrupted Scan Recovery

Progress is preserved through any interruption; continue or delete in one click.

Triple Validation

Triple Validation

Reference · satellite detail · DEM elevation model, side-by-side.

Live Notifications

Live Notifications

Instant alert for every high-score finding.

Candidate Pin Management

Candidate Pin Management

Approval/rejection workflow, filtering and KML export.

// PIPELINE

How it works

The lifecycle is summarized in four steps; behind the scenes, vision-language models, classical computer vision and geospatial analysis work in concert.

1

1. Reference Training

Multi-channel feature vectors are extracted from images of known archaeological structures (color, texture, edge, semantic embedding).

2

2. Boundary Definition

The area of interest is defined via KML or polygon; the system slices it into millions of meter-scale cells.

3

3. Satellite Scan

Each cell's high-resolution satellite image is fetched, compared with references, and a confidence score is produced.

4

4. Candidate Validation

Cells above threshold are pinned; experts validate with satellite imagery, DEM elevation model and matched reference.

// CAPACITY

Proven capacity in the field

1.484
Trained reference structures
5
Structure categories (castle, church, monastery, ruin, mound)
~0.25
Cells analyzed per second
0.65
Default confidence threshold
// PARTNERSHIP

Lead Investor & Strategic Partner Wanted

ARGEOS is currently in early investment stage. See our strategic partnership page for details.

Request a Meeting
// FOUNDATIONS

Scientific Foundation

ARGEOS combines techniques from the open literature into a single production pipeline. Which models we use is proprietary; how we combine them is scientific.

Multi-channel Feature Extraction

Histogram, texture (LBP), shape (HOG), keypoint descriptors and modern deep-learning visual embeddings work together.

Geospatial Index & Tiled Mosaics

A dynamic tile cache operating in Web Mercator projection serves high-resolution imagery on a per-cell basis.

DEM Topographic Analysis

Open elevation data (Terrarium/SRTM/ASTER) is visualized around the pin with hillshade and a color map.

NMS & Score Thresholding

Non-Maximum Suppression removes duplicate detections from adjacent cells; only the best candidate reaches the expert.

Ready to explore your region?

We evaluate collaboration requests from universities, museums, excavation teams and national institutions. Apply for early access.

Application Form
This site uses cookies to enhance your experience.