Case Studies / RapidLab
IoT · COMPUTER VISION & EDGE AI
Replacing manual headcounts with AI-powered people counting on edge devices
Buildings and public transport operators needed to track occupancy in real time – for safety compliance, capacity management, and operational planning. Manual counting was impractical across multiple entrances, and physical counting gates were too expensive or architecturally impossible. Working with RapidLab, we built compact AI-powered devices that count people automatically using sensor data and computer vision, processing everything on-device with no cloud dependency.
Client: RapidLab – a specialist in intelligent electronics and IoT prototyping, serving building operators and public transport providers.
KEY RESULTS
Edge AI
All processing on-device — no cloud, no internet required
2 modes
Time-of-flight sensor ML and computer vision deep learning
Real-time
Live occupancy tracking across all entrances and exits
Compact
Small-form-factor device for buildings and vehicles
INDUSTRY
IoT / Building & Transit Operations
USE CASE
Automated people counting
AI APPROACH
Time-series ML + computer vision
HARDWARE
Time-of-flight sensors + video cameras
DEPLOYMENT
Edge computing (on-device)
ENVIRONMENT
Buildings, buses, trains
The challenge
Knowing how many people are inside a building or vehicle at any given moment sounds simple – but in practice, it’s a hard problem. Buildings have multiple entrances and exits. Buses and trains have doors opening simultaneously at every stop. Manual counting is impractical at any meaningful scale, and physical counting gates – turnstiles, barrier systems – are expensive to install and often impossible in existing architectural layouts.
The need became acute during the pandemic, when occupancy limits and social distancing requirements had to be enforced in real time. But the underlying problem is permanent: building operators, transit authorities, and facility managers need accurate occupancy data for safety compliance, capacity planning, and operational efficiency – and manual processes can’t deliver it.
The core problem: real-time occupancy tracking was either manual (impractical at scale), gate-based (expensive and architecturally constrained), or nonexistent. Operators were making decisions about capacity and safety without reliable data.
What we built
Working with RapidLab’s IoT hardware engineers, we developed the AI layer for a compact device that automatically counts people entering and leaving a space – deployed at doorways in buildings or vehicle entrances on buses and trains.
Two complementary AI approaches. We built two versions of the counting system to handle different deployment scenarios. The first uses time-of-flight sensor data with machine learning on time-series patterns – lightweight and effective for simpler environments. The second uses video cameras with deep learning computer vision – more robust in complex, high-traffic scenarios where multiple people pass simultaneously or lighting conditions vary.
Edge AI – no cloud required. All AI inference runs directly on the device. No video or sensor data is transmitted to the cloud. This was a deliberate design choice driven by three requirements: real-time responsiveness (no network latency), deployment independence (works in locations with no internet connectivity), and data privacy (no personal video footage leaves the device).
Rapid prototyping with neural network architectures. The project involved defining data collection protocols, exploring sensor data patterns, and rapidly iterating through multiple neural network architectures and ML approaches to find the right balance between accuracy and the computational constraints of edge hardware.
Compact, scalable form factor. The end product is a small device that can be mounted at any doorway – no construction work, no turnstiles, no architectural modifications. This makes fleet-wide deployment economically viable across hundreds of locations or vehicles.
The results
BEFORE
Manual headcounts, expensive physical gates, or no occupancy data at all. No real-time visibility into how many people are in a space at any given moment.
AFTER
Automated, real-time people counting on compact edge devices. No cloud dependency. No architectural modifications. Deployable across buildings and public transport fleets.
The solution gave building operators and transit providers accurate, real-time occupancy data without the cost and constraints of physical counting infrastructure. Counting happens continuously and automatically – no staff involvement, no internet connection required, no privacy concerns from video leaving the device.
Beyond the immediate safety and compliance use case, the occupancy data feeds into operational planning – understanding peak usage patterns, optimizing staffing, and making capacity decisions based on actual data rather than estimates
Technology used
More Case Studies
See how we’ve helped other companies
AUTOMOTIVE (UNDER NDA)
Replacing reactive maintenance with a predictive maintenance roadmap
Unplanned equipment failures were causing costly production stops. We assessed the manufacturer’s infrastructure and delivered a concrete architecture for predictive maintenance, a roadmap from reactive repairs to data-driven prevention.
PdM
Architecture
Roadmap
Delivered
MANUFACTURING · RADAWAY
Making email-based order processing reliable with LLMs
Staff were manually reading customer emails, identifying products, and entering orders by hand. We turned a promising AI prototype into a production system that handles it end to end, across languages, formats, and attachments.
-90%
Manual intervention
95%+
Match accuracy
LOGISTICS · FR. MEYER’S SOHN
Eliminating manual data extraction from thousands of daily shipping emails
Operations staff were manually reading German and English logistics emails to pull out routing and scheduling data, every single day. We built an AI pipeline that extracts, structures, and delivers the data automatically.
–80%
Manual effort
2 langs
DE & EN
On-prem
Deployed