New Intelligence By iDter

How Video Analytics and Artificial Intelligence Will Revolutionize Remote Video Monitoring (RVM)

Harmonious integration of intrusion detection and automated deterrence will come from a single, vertically integrated platform manufacturer

Artificial Intelligence (AI) is a broad term that encompasses a wide swath of technologies aimed at solving problems. It’s used to automate processes and cull data for analysis, and is increasingly deployed in healthcare and medical diagnosis.

In the security industry, the primary use of AI has been in conjunction with video cameras to classify images for use in detection, facial recognition or the presence of guns or weapons. AI enables the development of sophisticated, targeted applications to improve the effectiveness of human monitoring labor and automate security detection.

CNNs or Convolutional Neural Networks are a category of neural networks used for this type of image classification and recognition. CNNs have proven successful in identifying objects, signs and even faces. These deep-learning algorithms take a series of images as the input to detect and assign importance to the various features of the image to differentiate one image from the other.

The security industry is now experiencing substantial growth in the remote video monitoring (RVM) segment through the deployment of cameras, AI input filtering and human monitoring personnel to protect a client’s property from loitering, vandalism and theft. The labor cost component to accomplish sustainable RVM results remains high. Missed detections and periodically overwhelmed operators continue to blemish the RVM space.

Automation in the RVM space ensures successful detection

The question really becomes how to automate the mundane tasks that distract operators from incidents and focus them only on intrusions that require human intervention. Humans should only be interrupted when it is determined by the system that an event demands human attention and engagement. CNN technology is only one component of the enablers that can be leveraged to harvest the full value of highly automated video monitoring.

The solution requires innovation singularly focused on the mission to successfully automate the RVM value proposition. Innovation in AI will continue to improve this value proposition. AI will become more sophisticated and fewer hardware and devices will be necessary to achieve complete, accurate detection in a single technology. The use of AI extends beyond improvements in detection and classification to uses that optimize automated deterrence behaviors and ensure they were effective and manage even unpredictable events successfully.

A multi-disciplined hardware and software team aligning a collection of technologies aimed at fully automating the detection, classification, immediate deterrence, recording, notification, communication, tagging and dispatch functions is now possible. The tools at our disposal to optimize RVM:

Quadcore system on chip (SOC) used to encode video, camera hardware design that combines video analytics and AI for fast detection and executing configurable deterrence behaviors.
RVM camera hardware design incorporating the tools to accomplish immediate deterrence using integrated floodlights, strobes, amplifier/speakers for alarms and voice recordings and a microphone.
Leveraging video analytics within the SOC to automatically zoom into areas of movement for extended distance coverage.
Training of CNN algorithms specifically for the RVM camera to reduce false positives to the lowest levels without ever missing a detection and classification of successful deterrence events to minimize human distraction from automatically resolved situations.
Integrated edge storage and cloud connectivity to host recordings and send push notifications.
iOS and Android applications that offer ubiquitous access to events, site controls, configuration and scheduling of the security system.
Advanced security applications running in the cloud that enable unpredictable deterrence responses and orchestration of multiple devices that leave no safe place for intruders to hide.
Integration with a cloud-based monitoring platform that graphically focuses operators on events that require intervention and police dispatch.

The companies best suited to accomplish this level of harmonious integration will be those who have the hardware and software development resources to take single-source responsibility for the development of the solution from end-to-end.

Effective RVM requires a seamless integration between the camera/lens design, use of the encoder SOC to auto-zoom images and extend the detection range and collection of large data sets of images for one specific camera. AI must be trained to assess all situations to avoid missed detections and false alarms. It must be programmed with highly effective deterrence behaviors that defy predictability; host event recordings and advanced security applications in the cloud; send timely event notifications to users; and optimize the human monitoring interface to focus operators on events that require attention. A cobbled solution that stitches together multi-vendor components and technologies is less likely to realize the same performance levels and come at a much higher operational cost.

The payout for innovative optimization and automation of the RVM business will be better security results, an order of magnitude reduction in monitoring labor cost, higher margins for the service providers, lower security costs for the clients and a greater adoption rate for RVM services.

White paper author:

Greg Ayres, VP Marketing & Business Development at iDter

August 13, 2024