Terrorist attacks on large crowds are a reality security professionals need to be prepared for. Deep learning-based video surveillance technology can help counter these types of threats.
Terrorist attacks on large crowds are a reality security professionals need to be prepared for. Deep learning-based video surveillance technology can help counter these types of threats.
Advanced surveillance camera technology can be used to investigate suspicious behavior and prevent terrorist attacks on public places.
The U.S. Department of Homeland Security lists some out-of-place scenes that security cameras should be set to identify: vehicles parked near areas with high pedestrian traffic, individuals wearing unseasonably bulky clothes and individuals attempting to enter restricted areas or entrances as patrons are leaving.
After receiving a bomb threat, the standard security practice is to initiate a search of the entire event space to locate the potential explosive device. If a suspicious package is detected, event operators can use camera footage to determine if the package has been deliberately placed or is merely unattended.
Security professionals are realizing
the importance of advanced video surveillance. IHS Markit figures show global shipments of network camera reached 108 million in 2018. IHS Markit estimates sales will go beyond 120 million in 2019 and 140 million in 2020. The share of high-megapixel cameras is expected to increase as well.
“As the number of camera shipments increases, a better way of analyzing and assessing the captured video is needed,” said Oliver Philippou, Research Manager at IHS, in a report.
As conventional methods fall short of enabling the needed analysis, deep learning - which mimics the human brain to learn from and understand the world through a hierarchy of concepts - offers a potential solution.
Deep learning allows security personnel to intelligently search through large amounts of recorded video for a crucial piece of information that may be contained in only a few frames. “In a police investigation, this has the potential to save hundreds of hours of officers’ time. Searchable analytics are therefore penetrating markets such as city surveillance, which historically searched video manually,” IHS’ Philippou said.
BriefCam is one such company using deep learning to provide enhanced video analytics. BriefCam’s video technology can distinguish between people and objects on the basis of different factors. “Security personnel must be able to quickly review and understand incidents that have occurred and rapidly locate suspects and perpetrators,” said Stephanie Weagle, CMO of BriefCam.
The company offered robust video analytics, allowing security personnel to “rapidly review video across multiple cameras and pinpoint people and vehicles,” Weagle said.
Security personnel can search suspects in the footage based on attributes like gender, clothing and direction and speed of walking.
With the help of deep neural networks and computer vision technologies, moving objects, like someone riding a bicycle, can also be separated and tracked.