AI supported video smoke detection

Smoke Finder is an edge device capable of capturing an image from any camera, enhanced with powerful AI algorithms, designed for authonomus smoke detection. SmokeFinder is now available to order, contact us!

Neuralbit Technologies sp. z o. o. is realizing a project co-financed by European Funds "SmokeFinder - an innovative forest fires visual detection system, empowered by artificial intelligence algorithms". Project main goal is to develop early smoke detection system with use of video surveillance systems. The innovation in approach to the matter is decetralized operation of machine learning algorithms which analyse images from surveillance cameras. The end device empowered with machine learning accelerator work directly at the data source - vision camera. This approach enables acces to highest quality video streaming which provide the better performance of algorithms. Due to that significant reduction of requirements for ICT structure neccesary for video streaming make the solution economically viable. The project's main outcome is complete automatic smoke detection system. By applying distributed computing, the system benefit from its accessibility and versatility. SmokeFinder can be integrated by plug and play manner with any existing video surveillance system. Project value: PLN 1 768 772.06, contribution from European Funds: PLN 1 332 157.35

Supported by

Created by

Effects of wildfires

Deforestation

Our planet is consumed by over 100,000 wildfires every year, which reduce forest area up to 170 million hectares.

Deforestation

Our planet is consumed by over 100,000 wildfires every year, which reduce forest area up to 170 million hectares.

Losses

Destruction of 1 ha of a forest by the fire is tantamount to lose amount wood material, which is worth about more than 5 000 dollars. However, indirect losses (related to non-productive functions of the forest and losses in the ecosystem) can be even 6 times higher.

Losses

Climate

One tree can absorb about 6-7 kg of CO2 from the atmosphere yearly. One ha of forest area offsets exhaust emission of 4 passenger cars during one year. One ha of forest area counteract annual exhaust emission of 4 passenger cars.

Climate

Solution advantages

Speed

SmokeFinder software is able to detect even a small amount of smoke in the image from observation cameras, and inform about a fire at a very early stage of its development. SmokeFinder is a great support for operators of a fire detection system - it does not get tired, does not distract, it works all the time.

Speed

Profitability

SmokeFinder can work with any already existing camera system, upgrading it with an intelligent image processing layer. SmokeFinder system does not require costly investments in the modernization of the ICT infrastructure.

Profitability

Precision

SmokeFinder enables a very precise mark of smoke observation angle, down to a fraction of a degree. Additionally, if SmokeFinder is installed on at least two watching towers covering the same area, it automatically determines the location of the fire.

Precision

Principle of working

Danger detection

SmokeFinder constantly analyzes images from observation cameras for smoke searching. If even small pillars of smoke are detected, frames representing the smoke are sent to the operator.

Data analysis

Advanced machine learning algorithms analyze the data collected from the observation camera in real time. If smoke is detected, its location is determined. Data from multiple cameras are combined to improve the firefighting action.

Services notification

If the threat is confirmed, the appropriate local fire brigade services are alerted automatically. The system transmits information about the location of the fire, its predicted size and additional forecasts about the direction and speed of the fire spread.

Fire fighting

By obtaining precise data from the network of sensors, foresters and fire brigades are better placed to adapt to the existing threat. In addition, SmokeFinder updates data on the fire development in real time, which allows for more efficient management of the firefighting action.