sferiacom@sferia.com  |  Privacy Policy |          
Home Page
Company Profile
Our teamv
Where we are

Via Albert Einstein 6
20090 Cusago (Milan Italy)

Phone: +39 02.4459180
Fax: +39 02.70048488

Home Page | Solutions | Other Systems | AUTOMATIC FIRE DETECTION
An AFD system receives images from cameras (pan-tilt-zoom), moving in a repeating pattern so as visit all the monitored area as frequently as possible. The repeating pattern can be changed to visit some areas more often than others, depending on pre-configured the risk and value assessments. The system can do local processing (using computers or digital signal processors near the cameras) or remote processing (using computers or signal processors far from the cameras at centralized locations). After the high-resolution high-quality images are acquired and corrected they are processed with a variety of algorithms to determine if there is a suspicion of fire or not. The best way to keep AFD sensitive at the same time as the false alert levels are kept low is by using user assisted learning, by constructing a database of images of the landscape at various times of the day and under various conditions, and then comparing each acquired image against this database and analyzing changes. One of the most important AFD feature is the ability to learn from user input. The users can instruct the AFD about circumstances that make a smoke plume a false alarm. The AFD automatically adapts itself to new circumstances by continuously updating its internal database to contain the most recent and relevant information. Adaptation allows the AFD to continue to operate properly when the environment changes (sun position changes over the course of a day or a season, tree leaf colors change, trees are cut, new roads are built, cars and other machines park in different areas, etc).
© 2018 TPSFERIACOM - Vat nr. IT 07064470151  
Certificate Number: 9160.SFRC
Certified Company UNI EN ISO 9001
Home | Company Profile | Our team | Solutions | Services | Contacts | Where we are | News