Products

PATTERN

PATTERN turns existing road cameras into a privacy-by-design AI system that helps cities improve safety, traffic flow and mobility planning.

PATTERN is an AI-powered, privacy-by-design decision-support system for urban mobility management. It transforms existing road cameras into intelligent sensors that analyse traffic flow, detect safety risks, and evaluate the impact of urban interventions in real time. By combining mobility analytics, automated incident detection, and a cloud-based dashboard, PATTERN helps cities improve road safety, reduce congestion, and plan evidence-based interventions while fully complying with regulations.

Greenroads' Solution

AI powered insights for safer, smarter cities. Instant and easy data monitoring and visualisations with any normal cameras or drone.

Greenroads offers cities and urban planners an understanding of movements on roads and open spaces via normal cameras. Using computer vision and machine learning, our algorithms can read a camera feed and convert it to GDPR compliant smart city data in real time. This information is easily accessible via a web browser and can be used to manage traffic flow, take data driven decisions, build digital twins, understand the impact of road works and plan livable cities, safer roads and low emission

MARVdash

Cloud dashboard for deploying and managing AI services, optimising resource allocation, monitoring, and scalable execution.

MARVdash is a cloud-based dashboard that manages the deployment and orchestration of containerised AI services. It provides a centralised web interface for configuring services, optimising resource allocation, and scheduling complex tasks such as video anomaly detection. Built with Python, and integrated with Grafana, Prometheus, and Loki, it runs on Kubernetes to enable scalable, automated, and efficient management of distributed cloud environments.

VideoAnonymization

Video processing tool that blurs faces and plates from video streams

The tool detect faces and plates in a video stream, blur them and provides a video stream in output with less personal information for further processing

VideoAnomaly

Anomaly detection in video streams based on vision language models

The tool leverages VLMs and LLMs to perform anomaly detection in video streams. Leverange the knowledge of the language models the tool is highly customizable and does not require anomaly-reataed videos

AudioAnony

The tool modify the voice in audio recordings to make the speaker not recognizable

The tool detects speech segments in audio stream and manipulate the speaker fingerprint to make it not recognizable while preserving the content and the prosody