A New Paradigm in Fire Safety with Digital Twins and Artificial Intelligence
Detectium revolutionizes anomaly detection and fire prevention for industrial facilities by integrating electro-optical/thermal (EO/IR) sensors technology with digital twinning and machine learning algorithms.
Our clients in the industrial and agricultural sectors benefit from our solution, utilizing it to proactively prevent fires while enabling the following features before, during, and after an incident:
- Anomaly detection to prevent fire
- Security integration allowing to evaluate real-time occupancy level
- Fire propagation scenario modelling and simulation
- Early fire detection
- Accurate fire localization
- Real time data feed available for civil defence responders “God’s eye view”
- Real-time escape route selection (based on evolving threat” for every floor.
- Root cause identification
- Damage assessment
- Insurance claim verification
We combine real-time 3D visualization with advanced sensors, and machine learning to create a digital twin solution for companies, which enables them to detect the fire much earlier than the conventional solutions.
Our clients in the industrial and agricultural sectors benefit from our solution, utilizing it to proactively prevent fires while enabling game-changing features before, during, and after an incident.
Product and Service Offerings
B2B - Industrial Uses
A digital twin is a real-time virtual representation of physical objects. Our solution comprises hardware and software components, cloud services, and machine learning (ML) technology, integrating Infrared and RGB sensors, 3D model visualization, and data analytics to elevate fire safety standards.
Our highly sensitive thermal and RGB sensors can detect even small flames indoors and outdoors. The operation of our flame detector is inherently much faster than that of conventional fire detectors based on smoke or heat detection.
We provide a flexible 3D visualization service tailored to the complexity and size of each site, catering to the diverse needs of our industrial customers. Our service can be utilized for a range of applications, including workflow simulation and optimization, as well as indoor navigation. With our advanced 3D visualization technology, industrial clients gain valuable insights into their sites, enabling them to optimize processes, streamline workflows, and enhance overall efficiency.
Leveraging historical data and predictive analysis, we offer comprehensive data analysis and machine learning solutions to foresee fire hazard risks at your facility. With our subscription-based computer vision service, you gain access to an ever-improving fire detection system that harnesses the power of advanced algorithms and cutting-edge technology. By subscribing to our service, you can stay ahead of potential fire hazards, proactively mitigate risks, and ensure the safety and security of your premises.
B2C - Household Uses
For our non-industrial customers, we offer a simplified version of our solution that harnesses the capabilities of computer vision. This streamlined offering provides an accessible and user-friendly experience, catering to a wide range of customers with diverse needs. By leveraging computer vision technology, our simplified solution still delivers effective fire detection and safety features, tailored to non-industrial settings.
The Benefits of Detectium's Solution
Our solution incorporates a state-of-the-art highly sensitive IR detector, enabling the detection of small temperature anomalies and flames both indoors and outdoors. In addition, we enhance the system with active learning on an RGB sensor, significantly elevating visibility and accuracy levels. With a rapid detection time of under 9 seconds for small fires, we outperform current fire detection systems, ensuring quick and efficient fire detection.
Our cutting-edge thermal and RGB camera sensor technologies offers an expansive detection range, setting a new standard in fire detection capabilities. Our solution stands out by detecting flames of 0.01 square meters size from an impressive distance of 60 meters away. In comparison, traditional fire detection methods using smoke detectors typically have a limited range of approximately 10.5 meters, and heat detectors come in at around 7.5 meters.
Use of data
By harnessing the power of machine learning algorithms to analyze historical data collected at the factory over an extended period, specifically focusing on various fire hazard incidents our solution intelligently identifies and maps high-risk areas, enabling more effective and targeted risk mitigation strategies for the future. Additionally, we continuously retrain our machine learning model to ensure its adaptability and efficiency in managing all possible scenarios.
The Benefits of Digital Twin for Safety
Our solution uses a highly sensitive thermal-RGB detectors capable of detecting small flames indoors and outdoors in less than 9 seconds. Early detection exponentially increases the fire extinguishment probability.
Our solution offers an expansive detection range, setting a new standard in fire detection capabilities by detecting flames of 0.01 square meters size from an impressive distance of 60 meters away.
Use of data
By harnessing the power of machine learning algorithms to analyze historical data collected at the factory over an extended period, specifically focusing on various fire hazard incidents our solution intelligently identifies and maps high-risk areas, enabling more effective and targeted risk mitigation strategies.
The beta release of our computer vision (CV) flame detector
Siavash is the CEO and co-founder of Detectium. He received a doctorial degree in Industrial Engineering and Management from Aalto University in 2018. He has conducted extensive research on digital twins. His vision is to create a world where the digital representation of all objects and entities can be efficiently and accurately created to enhance efficiency and safety in the physical world.
Alberto is a technical developer at Detectium, specializing in 3D modeling and design. He holds a master’s degree in Robotics and Artificial Intelligence from KTH and Aalto University. Alberto is responsible for 3D visualization and the technical aspects of the project. He has worked at companies such as Meta and Datadog, bringing valuable experience to the team.
Adriaan is a technical developer at Detectium. His wide range of expertise includes embedded software development, release and test automation, cloud infrastructure, and software development culture.
Pasi is currently the Chief Specialist of business development and innovation with LähiTapiola Insurance Company. He is Detectium’s industrial advisor who assists the company in driving growth and fostering innovation.
Professor Jan Holmström is currently a Professor of operations management with Aalto University, Finland. He is also an expert in supply chain management and design science research. With an extensive publication record, he is Detectium’s distinguished academic advisor.
Established in 2021, Detectium is a startup in Finland. The project results from a collaboration between three partners: Fenno-Aurum, TeamDev, and Aalto University, to reduce casualties from industrial fires by digital twinning technology and hi-tech sensors.
Fenno-Aurum is our launch customer and hardware supplier providing their established hardware experience and sensor innovations to continuously collect accurate data about the factory environment.
TeamDev is our project business champion who provides their expertise in machine learning, which is used to analyze the sensor data in real-me and yields actionable insights about improving the factory’s safety.
Aalto University carries over their proficiency in building simulation models from previous projects, which will manifest themselves as a digital twin of the factory. This digital twin will source its data from Fenno-Aurum’s sensors and TeamDev’s data analytics, integrating both into a virtual representation of the state of factory safety.
This project has received seed funding from the EIT Digital 2021 program as an innovation activity in the area of the digital industry under grant number 20644 (Project ID: 21325).
Digital twin definition and background:
A digital twin is the digital counterpart of a physical asset that specifically works with real-time data fed by sensor systems to record and analyze the real-time structural and environmental parameters of a physical asset. This is done for the purpose of performing highly accurate simulations and data analytics.