AI software makes cannabis business’ cameras systems smarter

Dec. 23, 2020
AI software can filter out 90% or more of alarms caused by nothing more threatening than a passing dog or blowing leaves

The legal cannabis industry offers many tempting criminal targets with its mix of quickly marketable products and often large amounts of on-site cash. That’s one reason states and local governments have set very high bars for security. Marijuana cultivators or retailers can’t cut corners without exposing their businesses to theft, vandalism, other crimes and the potential loss of a business permit.

With security regulations varying in locations across the U.S. and Canada – even between a city and its neighboring jurisdiction – it’s easy to see how business owners could become confused and frustrated. Integrators specializing in securing cannabis businesses have sprung up to help owners comply with regulations through detailed security plans, including access and inventory control, lights, vaults, gates and video surveillance. The quality and standard of surveillance video is a critical component of these systems. 

What Works in Video Surveillance

Properly placed, remotely viewed cameras are essential for both security and compliance issues. In most cases, surveillance cameras must provide an unobstructed view of anyone on the premises; all areas where marijuana products are handled; as well as outdoor perimeters and parking lots. Cameras at cultivation facilities must cover where marijuana is grown, cured, manufactured and waste is destroyed. It’s easy to see how the number of required cameras snowballs.

Recorded video is ideal for forensic purposes and providing visual proof regulations are met. But it’s the live feed that enables operators to see crimes in progress and immediately share data with security guards or local first responders. Some jurisdictions require live monitoring 24/7. However, video monitoring is not a DIY project; it’s better to leave it to professional central station operators.

It’s no secret within the security industry that 95% or more of traditional video alerts are false. There are too many unnecessary alarms added for already harried monitoring personnel. An average monitoring center operator sees three alarms per minute, each needing to be checked closely for criminal activity. Sorting through false alarms wastes time and genuine alerts may be missed entirely.

False alarms occur in virtually any environment. However, cannabis cultivation areas, mainly open-air and outdoor sites, are particularly prone to exposure to the elements. Artificial intelligence-based software provides a solution. More than a decade in development, AI software can filter out 90% or more of alarms caused by nothing more threatening than a passing dog or blowing leaves. With fewer alarms to review, operators have more time to spot actual alarms and follow protocol. 

AI-based Software Enhances Video

Here’s a brief look at how the time-consuming task of training the software’s neural networks pays off. AI-based software is exposed to millions of photos showing people of various ages, wearing different clothing and in many environments. Human trainers correct mistakes made by the software throughout the process. 

Over time, the software’s neural networks become increasingly accurate. The same process gets repeated with vehicles which are often involved in any break-ins and other crimes. An added benefit to the technology is that AI-based software continues getting “smarter” after installation.

The cloud-based software easily integrates with a monitoring station’s existing operating software. The cameras located in and around a cannabis facility are added through an easy drag-and-drop process. A software dashboard highlights idle cameras, enabling an integrator to see if they need replacing or should be moved to another area to increase their value.

Recent developments in AI-based software are bringing us closer to the extinction of false alarms. Yet, the industry has only begun to tap the full capabilities of artificial intelligence. Next is software to identify human behaviors associated with criminal activities – before a crime is committed.

The security requirements placed on cannabis operations are extensive and expensive. Ensuring a business’ monitoring service is using the latest AI-based software enables owners to feel confident video from their facilities isn’t awash with false alarms.

About the author:Brian Baker is vice president, Americas, Calipsa. Calipsa is a leading provider of video analytics for false alarm reduction