Video Surveillance Analytics for Corrections… Why Not?

By Lyle Cutler and Craig Clary

We get it. Video analytics applied to surveillance systems in correctional facilities still seems to be the mangy dog no one wants to pet; and rightfully so. Historically, analytics have been slow, difficult to deploy, required multiple servers, were a complete nuisance to maintain, and ultimately came with a hefty price tag. But like most technology innovations, it either improves in efficiency and cost or eventually fades away. Those in the security world are insisting on the creation of faster, more intuitive, and economical analytics at an alarming rate. Video surveillance manufacturers are responding by heavily investing in research and development as the demand increases. The use of analytics has become a part of daily life and it’s unlikely that will change in the near future.

With such a wide gamut of analytics available to the consumer, it can be difficult to determine what makes the most sense in a corrections environment. A long list of analytic technologies is available to the security market, but not all apply to a correctional setting. The following are a few scenarios where analytics may be used as mitigation to potential situational pain points.

Fence Lines

The perimeter fence is, of course, what keeps detainees…detained. In a typical correctional setting, we see a combination of perimeter fence shakers (perimeter detection) systems, below-ground shakers, microwave breech sensors, guards walking fence lines, and patrol vehicles driving around the perimeter for hours on end. With single and most basic video analytics, most if not all of these systems may be eliminated. As a rule of thumb, surveillance cameras are typically placed in strategic positions where the command center (master control) is given a complete view of the entire fence line. An object tracking analytic is applied by drawing a virtual line within the camera’s field of view to determine where an object is or is not permitted. When an object crosses this line, an alarm is created (both visually and audibly), to notify the user precisely what has generated the notification. Multiple virtual lines may be drawn to outline inner and outer fences. With the addition of “distance-to-line parameters,” thresholds may be created within the analytic providing specific areas of non-detection (masking) such as isolation zones or perimeter patrol roads. The analytic is also refined enough to determine what direction the object is traveling providing additional potential threat information to the user.


On average in the U.S., one physical altercation per facility per day occurs either between inmates or between an inmate and a corrections employee. Unfortunately, this average is on the rise. Early awareness of an incident is paramount to the de-escalation of a potential altercation. Two analytics stand out as extremely helpful in situations of this type. Crowd detection is an analytic mostly based on the ability to detect the number of people within any given area. In a corrections environment, physical violence and crowds are historically conjoined. The analytic may be directed to alert a user when a preprogrammed number of people are found gathered in a certain area. The analytic also provides a live count of the people within its view. Additionally, an analytic can utilize microphones, built into the camera or stand-alone, to detect shouting or elevated noises. Most physical altercations are preceded by verbal aggression or sound atypical to an averaged ambient noise level. Audio-based analytics have been developed with the ability to discern between the typical elevated conversation and a full-blown shouting match. These analytics may be used in any situation where aggressive behavior could be a precursor to violence. Be aware that federal, state, and local laws may not allow the use of microphones within a facility, so it’s important to consult local regulations and insurance carriers before implementing any microphones or audio analytic technologies.


Slips, falls, medical events, and physical altercations are only a few of the hazards and situations that may be found in any occupation. These incidents may be more likely in the corrections field due to the physical nature of the job. Examples include:

  • A guard on tour at 2 a.m. who has a heart attack while making their rounds and master control does not see this person on the floor.
  • A person is walking to their vehicle late at night and slips on the ice, hits their head, and is lying on the ground and no one is aware of the incident.
  • A guard has been assaulted by an inmate and master control missed the event.

Posturing analytics possess the ability to determine when a person is standing, crouching, sitting, or lying down and provide alarms/notifications to users based on parameters programmed to the analytics, mostly centered around the amount of time a person holds any given posture. Think back to the aforementioned ice incident, a threshold of time is needed for the person to recover and continue walking. Otherwise, the analytic would provide a false alarm, and security personnel would unnecessarily be pulled away from their assigned duties. Conversely, a person in a crouching posture could likely be preparing for a physical altercation and guards should be alerted promptly should the analytic detect such a sustained position.

Visitor Awareness

Utilizing analytics can be hugely impactful for securing the premises of a corrections facility. Because there are certain individuals who are not authorized to be on the property, early detection of these situations may be found with facial recognition and license plate recognition analytics. Both analytics are being further refined with greater success. Cameras may be placed at parking lot entrances viewing daily traffic while the analytic is in the background sending license plate information to a registry and, in turn, reporting issues to master control. This same scenario may be replicated by placing cameras at door entry points allowing the analytic to send facial information to a registry and provide alarms and notifications as necessary.

As manufacturers move away from server to camera-based analytics, the cost for this service has been greatly reduced since its inception. Camera-based analytics are also cheaper to maintain, as server hardware and upgrades are removed. Server and cloud-based analytic services are also an option if a facility is utilizing newer analytic compatible cameras and chooses to maximize its technology investment. 

Could the use of video analytics replace guards or other security personnel in a corrections facility? Absolutely not. However, it is important to recognize a single guard does not possess the ability to monitor 300 or more camera views at once. Analytics enhance the peripheral awareness of the security staff by allowing the analytics to monitor specific situations and the guard to focus on the main event areas of the facility.

When looking at ways to increase the safety of staff and the overall security of your facility, we encourage the consideration of video analytics as a valuable enhancement to facility security measures.

Lyle Cutler is a Senior Security Designer and a Senior Project Manager with Dewberry. Craig Clary is a Senior Security Designer and a Senior Project Manager with Dewberry. 

Editor’s note: This article originally appeared in the May/June 2023 issue of Correctional News.