Surveillance Cameras with Artificial Intelligence

Surveillance Cameras with Artificial Intelligence: morphology-based human and vehicle detection

The incorporation of morphology-based detection in CCTV cameras with artificial intelligence is crucial for minimizing false alarms triggered by irrelevant objects like animals loitering in the area or natural environmental elements. This technology ensures that alerts are exclusively generated for detections of interest, such as humans and vehicles.

The potential of artificial intelligence in CCTV security cameras

• Face detection and recognition: Face recognition can be used in a variety of applications. It enables the triggering of alarms upon detecting a suspicious face stored in a database. It can also be employed for tasks such as time and attendance control, facilitating the monitoring of staff’s working hours within a designated work area.
• People counting: in many shops it is essential to know how many people have entered and left the premises. In addition, people counting cameras offer statistical reports that will help the business with marketing and/or promotional actions.
• Video metadata: Any forensic search to identify a person or vehicle in the event of an offence requires time and dedication to find the suspect. Video metadata helps to be able to provide additional information when conducting a search. For example, it can be filtered to identify a person wearing a blue cap, backpack and blue clothing who is alleged to have robbed a shop in a shopping centre.
• License Plate Recognition (LPR): cameras with license plate recognition can be used for control at entrances and exits to trigger, if required, an access barrier. Similarly, LPR cameras are used for traffic control in smart cities to collaborate with road safety or to control traffic flow with periodic reports.
• Human and vehicle detection: Nowadays, equipping an installation with Artificial Intelligence is not complicated. The detection of humans and vehicles adds extra security to an installation, as they help to send an alarm to the client in the event of an intrusion in a restricted or private area.

How does the AI camera learning system work?

The artificial intelligence that is present in a CCTV system may be housed in the camera or in the recorder. Most systems equipped with artificial intelligence contain a database or trained libraries with many samples and times. These have been pre-approved so that the camera can compare and be certain that what it is seeing, at this precise moment, is a target with the appropriate size and morphology from the database.

What kind of end-customer will be able to benefit from this technology?

This function is valid for a wide variety of indoor and outdoor scenarios:

• Warehouses / Factories: With intrusion detection you can control a space to generate alarms with humans, so you can manage access control of people by zones.
• Parks / Pedestrian streets: With line crossings you can configure SIP to generate an alert when it detects a vehicle (cars, motorbikes...) inside a park or pedestrian area.
• Villas or detached houses: With line crossings and intrusion detection you can keep an eye on the perimeter area to detect people jumping, people in the garden or vehicles entering the garage.
• Public and private car parks: With intrusion detection you can control vehicle access ramps by generating an event and alarm when the presence of pedestrians is detected.

Safire Smart metadata technology: artificial intelligence in PTZ cameras

The morphology-based detection included in PTZ cameras is fundamental to reduce false alarms produced by objects that are not of interest (animals wandering around the area, elements of the natural environment, etc.), sending alerts only for human and vehicle detections.

• Autotracking: automatic tracking of the target as it travels through the premises. In contrast to fixed cameras, these PTZ devices offer motion capabilities: notably, intelligent autotracking is a compelling function that allows the position of the detected target in restricted areas not to be lost at any time. It is an automatic tracking that starts when the cameras detect a human or vehicle in an area of interest and tracks it as it moves through the installation until it disappears from the scene.
• Active Deterrence: cameras issuing pre-recorded message and deterrent flash. The new Safire Smart PTZ cameras feature active deterrence, which can broadcast a pre-recorded message and flash a deterrent flash when an intruder is present. This function can be triggered from its own smart event or from smart events from other Safire Smart cameras co-existing on the same recorder (for example, if a fixed camera detects a human, it can trigger an "event" to the PTZ camera to rotate and activate its speaker and deterrent light).
• Long-distance infrared illumination: IR Laser range up to 500m. A PTZ camera with laser IR up to 500m offers enhanced night vision (laser IR provides intense illumination in low-light environments or in complete darkness) and extended range (IR lasers have a significantly longer range compared to conventional infrared LEDs).

SIP (Smart Intrusion Prevention) function, the artificial intelligence of Uniview

Uniview's Smart Intrusion Prevention (SIP) is a perimeter protection technology with artificial intelligence. It is focused as a more professional and advanced security solution than other systems such as conventional motion detection.

Uniview features the Smart Intrusion Prevention (SIP) function on its Prime series camera. This AI technology allows 4 types of intelligent functions to be configured:

• Line crossing detection
• Intrusion detection
• Entrance area
• Exit area

HIKVISION AcuSense Technology:

The innovative AcuSense technology is based on an advanced deep learning algorithm developed by Hikvision's R&D team.

AcuSense technology excels in its ability to identify targets, accurately discriminating between people and vehicles from other moving elements (pets and weather elements such as rain).

AcuSense triggers an alarm only when the pre-defined type of intrusion (either people or vehicles) occurs. Thanks to accurate target classification, the system can filter out up to 90% of false alarm events, achieving a significant improvement in system efficiency and effectiveness:

• False alarm filtering.
• Strobe light and customisable audio alarm.
• Intelligent content search.

Thermal cameras with artificial intelligence

Sunell AI applied to thermal vision

Wildfires are a huge natural disaster with a devastating impact. To provide early warnings and prevent fire incidents, Sunell products actively employ artificial intelligence technology.

Through AI analysis and deep learning-based algorithms, Sunell's thermal imaging products can detect temperature anomalies in key areas and quickly trigger alerts to prevent fires. In fire-prone regions, thermal imaging cameras use intelligent fire point detection algorithms to detect fires in their early stages.