Särkänniemi is the second most popular amusement park in Finland, built on a 50,000 sqm area with about 1,100,000 annual visitors. Featuring an aquarium, planetarium, children’s zoo, art museum, and the iconic Näsinneula observation tower, maintaining a safe and fluid customer experience is a massive logistical challenge.
The park wanted to extend its existing surveillance system with data analytics and AI. The goal was to enable automatic data collection of customer flows and establish more efficient monitoring of security incidents. By accessing more accurate data, Särkänniemi aimed to optimize workforce management and respond more precisely to customer needs, ultimately improving both safety and the overall customer experience.
The Objectives
To deliver an intelligent security and analytics layer, the project focused on three core targets:
- Automatic Visitor Flow Data: Establishing a real-time count of visitors entering and leaving the park across multiple checkpoints, as well as crowd-counting the size of audiences at stage areas.
- Automated Security Monitoring: Detecting security incidents automatically, including monitoring vehicle gates for visitor safety and securing restricted areas against trespassing during closing hours.
- Utilizing Existing Infrastructure: Applying video analytics directly to the existing IP camera feeds within the park’s network, ensuring easy data usability through browser-based interfaces and business intelligence (BI) integrations.
The Solution: Non-Intrusive Edge Processing
The camera feeds from existing IP cameras at theme park entrances and gates were connected directly to the MarshallAI machine vision platform. To guarantee absolute data privacy and security, the platform was deployed as an edge service with the processing unit placed locally at the Särkänniemi HQ.
Local processing was deliberately chosen to ensure that proprietary data and surveillance camera footage would never leave the internal network. The MarshallAI service does not store any video footage or other sensitive data during usage. The data gathered is strictly numeric, limited to headcounts of people entering and exiting the park.
Successful Implementations & Results
The deployment proved that machine vision is an exceptionally versatile technology for monitoring and automation tasks. The project yielded immediate operational benefits:
- Dynamic Staff Placement: Real-time information on visitor flows enabled the park to deploy staff more efficiently and dynamically based on actual crowd density. Good levels of accuracy were achieved with minimal development time, even when utilizing sub-optimal legacy camera placements.
- Automated Off-Hours Security: Manual security monitoring was significantly reduced. The system was calibrated to detect trespassers during park closing hours, sending real-time automated alerts to security staff.
- Seamless Systems Integration: The platform was integrated into both Särkänniemi’s internal business intelligence platform and the Insta Blue Aware IoT-platform to maximize data usage. Because MarshallAI is natively compatible with many Video Management Systems (VMS), alerts and data can be transferred directly to systems like Milestone event servers.
“Video automation has proven to be very useful for our operations. This season we focus on integrations and better data usage, after which we will broaden the use of machine vision.”
— Miika Nenonen, Field Operations Manager
Looking forward, Särkänniemi and MarshallAI have identified multiple additional use cases for future implementation, including demographic detection and parking monitoring, features that are already supported off-the-shelf by the MarshallAI platform.
