CCTV video analytics provides businesses with a powerful method for recognising and detecting intruders, resolving crimes, keeping track of the movement of punters around their premises, and producing an alarm for specific types of antisocial activity. Most modern CCTV surveillance systems come with on-board video analytics, with advanced algorithms that provide advanced warning of criminal activity based on body language and movement patterns, and facial recognition software to aid the detection of known criminals.
As part of an integrated security system, a video analytics package can provide improved security to a business, deterring crime and improving safety for employees and customers alike. Video analytics also facilitates a more responsive and lighter touch approach to security, allowing your team to respond quickly to genuinely suspicious or criminal events, while keeping a less intrusive profile at other times.
There are several types of video analytics tools used in commercial CCTV surveillance.
1) Motion Tracking
When a moving object crosses a boundary, a motion tracking function detects it. This is accomplished by either drawing a virtual line across a specific area to create a ‘region of interest’ or by using the camera’s entire field of view as the boundary. In this case, any movement detected against the fixed background scene is considered a breach. Any predicted motion, such as birds flying or plant movement, can be discounted using filtering capabilities.
2) Object Detection
Object detection algorithms are configured to identify the presence of objects (in general) within the field of vision, whether in motion or not. This could, for example, be programmed to detect an unexpected object – such as a handbag left unattended in the games room – or the presence of a person in a restricted area.
3) Object Classification
Object classification goes one step further than object detection, by identifying the type of object within the field of vision. This function can serve a range of purposes, such as detecting and categorising the presence of specific weapons. Object classification can be linked to alerts and alarms if certain objects are detected.
4) Direction Flow
This video analytics function predicts the direction of motion based on the behavioural patterns of a moving object, such as a person or vehicle. This can alert your security team to potential security issues before they occur.
5) Loitering Detection
Loitering detection tools flag up unusual lack of motion around suspicious areas, such as a person loitering around a gaming table or strongroom with no discernible reason. Individual pedestrians inside the coverage area can be identified and tracked, recognising individual people by facial features and by the colour of their clothing in subsequent frames. Even if the pedestrian is no longer visible to the camera, the analytics information can re-identify him or her when they return to the coverage area or move into the field of vision of another camera.
6) Removed Object Detection
Removed object detection identifies when an expected object is no longer present in the field of coverage and can alert security personnel to potential thefts. These objects can be recognised and tracked around the premises. The sensitivity of this tool can be adjusted to meet your requirements, such as configuring alarms to only sound when an object is removed for a predetermined amount of time. This avoids false alarms being triggered when objects are moved for legitimate reasons, or transferred between rooms.
7) Facial Detection And Recognition
Video analytics software can detect individual faces in colour images using a range of algorithms. Some cameras are programmed to detect faces in general – which is facial detection: these cameras will alert you to the presence of human faces in an area – and others can be set up to link facial features to specific human beings, which is facial recognition.
8) Head Counting
Head or people counting software identifies the movement of individual people in and out of the field of coverage. This is useful for crowd control and evacuations in busy areas (e.g. games rooms or bars) and gives a more responsive and accurate picture of usage than manual headcounts. Some people counting software is very sophisticated, identifying the gender and approximate age of people in each area, helping businesses understand their user demographic and peak service times.
Next Steps
At HD Integrated, we supply advanced CCTV and video analytics systems for businesses in the retail, leisure, and night-time entertainment sectors. To find out more about the potential of video analytics for improving safety and improving the customer experience in your business, please call one of our security specialists today.
Image source: Pixabay