Improving Efficiency in Warehouse Security and Operations

The Operational and Security Benefits of Video Intelligence Software  

Managing a production facility involves overseeing such areas as logistics, human resources, and operations, as well as the security teams tasked with protecting and securing buildings, inventories, employees, and visitors. Given the size and complexity of some warehouses, their large inventories, and long working hours (many of them operate 24/7), the operational security demands of warehouse managers are multitudinous and ever evolving.  

While most warehouses already have video surveillance systems, some have also implemented complementary AI-driven video intelligence software to translate live or archived video into structured data. Based on Deep Learning and artificial intelligence technologies, video intelligence software conducts complex object extraction, recognition, classification, and indexing activities, so that the rich video metadata can drive crucial analytics, making video searchable, actionable and quantifiable: This enables security teams to: 

• Review footage from past incidents in a matter of minutes rather than hours with comprehensive and precise video filtering 
• Increase situational awareness and response time to evolving situations with rule-based, real-time alerts 

• Helps inform strategies for preventing future operational or security problems by visualizing video data so operators can uncover trends and patterns

In a previous blog post we discussed the health and safety benefits of video content analytics and surveillance in warehouses and production facilities; this post describes the operational and security benefits of video intelligence software in such work environments. Warehouses and production facilities have multiple administrative offices, loading docks, maintenance zones, storage and parking areas that require oversight, either in-person or via video monitoring. In implementing technologies, managers typically are concerned with ensuring perimeter security; reducing the need for security personnel to physically patrol (thus protecting their safety); and preventing future and investigating past incidents of inventory theft. For operations and security managers seeking to streamline logistics and maximize efficiency while preventing, responding to and investigating incidents, intelligent video surveillance offers compelling solutions to everyday challenges:  

Rapidly Reviewing Incident Footage 

Obviously, surveillance cameras deter theft – but only if perpetrators notice them. Inventory theft is still a common occurrence, even with extensive on-site video surveillance. Post-incident investigations invariably include reviewing relevant video footage, but manually combing through video is time-consuming, and human observation is prone to error. Because video analytics software has powerful search and filter capabilities, it drastically reduces the amount of time it takes to review incident footage; it can be reviewed in a matter of minutes rather than hours or days. 

Real-Time Alerts Improve Situational Awareness and Response Time 

Security managers need real-time, actionable intelligence: what is happening now, and where. A video analytics system can be configured to trigger alerts for anomalous behavior in a scene, based on rules set by the operator. Real-time video analytic alert triggers range from:  

• Unexpected lighting changes at pre-set times of day or night 
• People or object counts exceeding a pre-defined threshold 
• Dwelling in a scene for longer than normal benchmarks 
• Facial recognition of specific individuals on a suspect watchlist or of persons not included in a watchlist of authorized personnel 
• License plate recognition for vehicles included or excluded in a watchlist

For example, if a light is turned on or off during a pre-selected time frame – an unusual time for the facility – an alert to management can be triggered, allowing for immediate assessment, decision-making and response. This is especially critical for protecting workers in warehouse or factory areas that may be sensitive to human occupancy. Even though security staff cannot physically patrol those areas, they are still required to manage and maintain the integrity of these areas and ensure that people are not accidentally walking into harm’s way.  

In addition, video analytics system can detect whether workers are wearing protective gear, such as hard hats for safety in pre-defined areas, or – during a public health crisis –whether workers are wearing face masks.  

People-counting alerts can track and manage occupancy throughout the facility; for safety reasons, there may be a limit to the number of people who can be in a loading area, for example. Or, the system can be configured to detect humans in an area that is normally off-limits to people.   

Video content analytics can also detect dwelling in a particular camera view, based on pre-set rules; the operator of the system can set a time threshold to detect abnormally long lingering of people or objects and, in this way, respond proactively to potential problems, such as a suspicious package left on the property or a person who may be intending to commit a crime or having a medical emergency. This can be useful for securing the facility perimeter, by triggering alerts for dwelling or activity when the factory is closed to workers.   

Facial recognition alerts can be set up so that managers are notified when a face in a video camera feed matches that of a digital video or still image of a face that has been uploaded to a watchlist. This could be useful to alert management when a suspicious individual tries to enter the warehouse or production facility; for example, management may want to create a watchlist of former employees who are disgruntled and pose a potential threat. Conversely, a digital image watchlist can also be used to allow access to certain facilities; approved employees will be allowed access only if their face matches one of the approved faces on the watchlist.  

Managers of a facility might also apply license plate recognition functionality. For example, if unauthorized vehicles excluded from an employee or approved visitor watchlist are detected on the grounds, the surveillance operator can be alerted, assess the situation, monitor the vehicle, and respond if necessary.  

 Improving Operational Efficiencies with Business Intelligence 

Among their many responsibilities, operations managers must keep track of the stock inventory within their warehouses or production facilities, which is usually managed via pick order forms and processes. Occasionally, stock is moved without following proper procedures, and can be misplaced. The ability to search video quickly and pinpoint objects of interest can help managers quickly investigate and understand an occurrence to locate missing items and rule out theft. Forensic analysis can easily identify who last had the item, where, and when, for tracing the stock across multiple cameras to its final destination.  

Another advantage of video analytics is that it aggregates video data over time, so system operators can create customized reports and visual dashboards, including graphs, heatmaps, and histograms for demonstrating patterns and trends. Commonly, management teams want to see which loading docks have the most activity, or which aisles or floors of a warehouse or facility have the most vehicular, machine, or pedestrian traffic. Given that information, they can understand which products or stock items are most popular, or which areas of a warehouse experience traffic congestion and bottlenecks. Understanding these trends in a quantitative way helps managers plan their warehouse layout and stocking and make better decisions that are based on statistics rather than human observation. Optimizing traffic within a facility increases employee productivity, protects workers against injury and ensures streamlined customer service and fulfillment. And, this information can be collected across many warehouses or production facilities — locally, nationally, and internationally — to improve processes company-wide.  

Facilitating compliance with COVID-19 masking and social distancing mandates 

As I mentioned in my earlier blog post about health and safety, the COVID-19 pandemic is an unprecedented and difficult challenge for any workplace, and that is certainly true for warehouse or production facilities, because they employ hundreds, or even thousands of people. To ensure workplace health and safety, businesses need to deploy new capabilities to maximize compliance with evolving mandates, including social distancing and mask-wearing.   

Video intelligence software can play an important role in those efforts, because it streamlines contact tracing by enabling production facilities to forensically review video data to identify and trace diagnosed individuals. By using Face Recognition and Appearance Similarity, surveillance operators can easily search video footage to discover if a self-identified diagnosed person has been in contact with others on premises and, if so, with whom and for what duration of time. Employers can still protect the anonymity of the person who is diagnosed, but the people who have been exposed can be notified so they may then choose to self-quarantine to protect others from possible infection. The technology can also rule out those who have not been in close proximity to a diagnosed person, so they can be more at ease and will not have to self-quarantine.  

With proximity identification capabilities, video analytics software can also quantify and analyze distance between individuals over time and location to identify non-compliance with physical social distancing mandates. This can help managers gather anonymous statistical data about whether employees are generally complying with a rule to stay at least six feet apart from each other and make intelligent decisions about improving compliance to maintain health and safety in the factory or warehouse.  

Similarly, the system can count the number of people in an area, and alert staff in real-time if a crowd is forming in one part of the facility. Or, if managers prefer to not receive real-time alerts, they can simply review the data that the video analytics system aggregates over time, so they can have quantitative trend data about behavioral and pedestrian traffic in the facility.  

Facilities and warehouses clearly get more out of their existing video surveillance when they complement these systems with video content analytics for video forensic search, real-time alerts and operational intelligence.