GeoForensics AI — AI-Powered Image Geolocation & Forensic Intelligence
This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle
What I Built
GeoForensics AI is an open-source forensic intelligence platform that attempts to determine where a photograph was taken using visual evidence — even when the image contains no GPS metadata, street signs, or obvious location clues.
Instead of simply predicting a country or city, GeoForensics AI analyzes the image like a digital investigator.
It examines things such as:
Architecture and buildings
Roads and road infrastructure
Terrain and vegetation
Vehicles and road markings
Signs and visible text
Urban layout
Landmarks
Environmental characteristics
Visual clues that can be matched against geographic data
The system then combines these signals to generate a micro-location hypothesis and presents the result on a map with supporting evidence.
The potential users are OSINT investigators, journalists, researchers, security teams, insurance investigators, and anyone working with image-based geolocation.
The long-term goal is to turn this into a practical visual geolocation and forensic investigation platform, rather than another black-box image classifier.
Demo
GitHub: GeoForensics AI
Code
The entire project is open source:
GitHub Repository:
https://github.com/abdullahkapadia/GeoForensics
You can inspect the architecture, experiment with the system, and build on top of it.
How I Built It
GeoForensics AI is built around an AI-powered visual analysis pipeline.
The basic workflow is:
Image → Visual Understanding → Evidence Extraction → Geographic Reasoning → Candidate Locations → Map Verification → Final Location
The AI analyzes the uploaded image and extracts potentially useful geographic signals.
Instead of trusting a single prediction, the system is designed around evidence aggregation.
For example:
Architecture suggests Southern Europe
Road markings suggest a specific country
Terrain narrows the region
Building style suggests a particular city
Map/geographic data is then used to validate the hypothesis
This makes the system much closer to a forensic investigation workflow than a conventional image classifier.
Why Does Open Innovation Matter?
Geolocation is a problem where relying on a single closed API can quickly become a limitation.
Open-source AI makes it possible to experiment with different models, combine multiple sources of evidence, run components locally, and build a transparent reasoning pipeline around the models.
More importantly, the project itself can evolve through community contributions.
Someone can improve:
Visual clue extraction
Geographic reasoning
Map integration
Candidate ranking
OCR
Image analysis
New open-weight models
Evidence visualization
Instead of having a black box that simply says “I think this is Paris”, the goal is to build a system where developers can inspect, improve, and extend how the conclusion was reached.
That's what makes open innovation particularly valuable for this project.
Prize Categories
Open-Source AI / AI Application
GeoForensics AI combines open-source AI with geographic intelligence to create a practical forensic investigation workflow.
Why I built this
Most AI geolocation systems stop at:
“This image is probably from France.”
I wanted to ask a harder question:
“Can AI investigate the image and narrow it down to the actual street?”
That question became GeoForensics AI.
The project is still evolving, and the next step is making the system better at combining AI + maps + geographic databases + visual evidence into one investigation workflow.
Top comments (0)