[ Vision AI ]
01 - Vision AI
Computer vision trained on African road contexts. Detects road users, classifies behaviour, and flags risk events in real time across the network.
Our custom-trained Vision AI model monitors how vehicles, motorcycles, and pedestrians move through roads and junctions in real time, revealing movement patterns, safety compliance, and emerging risks. Built for African urban conditions: dense traffic, informal road users, mixed mobility, and variable infrastructure.
Vehicle types: motorcycles, cars, trucks, tuk-tuks/rickshaws, pedestrians
Estimated approach speeds and spacing
Helmet use and rider passenger counts
Road crashes are not just caused by driver behaviour, they are shaped by what the road looks, feels, and functions like. Our infrastructure assessment layer evaluates road design, signage, markings, junction geometry, and surrounding conditions.
Road condition, markings, and signage visibility
Junction geometry and sight-line analysis
Speed-management features: humps, signage, lane narrowing
The third layer of Mobility Intelligence connects road behaviour and infrastructure evidence to what is actually happening to people. In collaboration with Tamale Teaching Hospital (GHS), we integrate non-identifiable RTI data to identify crash hotspots.
Location-based crash hotspot mapping
Severity and injury pattern analysis by corridor
Direct linkage between behaviour and outcomes
In response, ILMIHA Labs evolved this pilot into a broader Mobility Intelligence platform, combining Vision AI, behavioural observation, and infrastructure analysis to generate scalable insights for policy, enforcement, and safer transport system design.
Three integrated layers, built from street-level observation in Tamale. Together they form a single evidence base for road safety decisions.
Computer vision trained on African road contexts. Detects road users, classifies behaviour, and flags risk events in real time across the network.
Physical asset mapping and condition scoring - junction geometry, signage and crossings - tied to where risk actually concentrates.
Injury records linked to location, time and road type, in collaboration with Tamale Teaching Hospital.
Behaviour, infrastructure and injury outcomes, unified - so cities can target interventions where they matter most.
Implementation partner. Provides crossing inventory, institutional authority, and public-sector deployment route.
Health data partner. Provides non-identifiable RTI data to link injury outcomes to road location and behaviour.
Strategic uptake and scale partner. Road User Intelligence data shared via platform for national oversight.
Engaged on evidence-based policy for motorcycle formalisation and school-zone safety infrastructure.
Transforming safer roads with a focus on school zones in Ghana. We've mapped over 81 points in Tamale, combining Vision AI speed analysis with clinical health outcomes to prioritize urgent crossing redesigns where risk is highest.
Read Case StudyA living evidence layer for African secondary cities. We surface real-time behavioral patterns, helmet compliance, passenger loading, and high-risk maneuvers, enabling agencies to shift from manual counts to continuous data-driven enforcement.
Open the RTI FormScaling street observation into evidence-based policy and transformative infrastructure across Ghana.