Video analytics for factories has advanced quickly in the past few years with the onset of edge-based AI, powerful processor architecture, and industry-standard protocols.
Factory video analytics is an extremely important area for physical security companies and systems integrators. Video analytics has advanced quickly in the past few years with the onset of edge-based AI, powerful processor architecture, and industry-standard protocols. Using them on factory floors with existing security systems has revolutionized manufacturing processes.
“This has allowed smart cameras to provide a new way of monitoring and controlling safety measures and optimizing process flow,” explains Richard Gibbs, Senior Applications Engineer at Honeywell. “It has also led to object classification-based algorithms updating building automation systems when humans enter a zone for safety purposes, and behavioral algorithms help detect worker behavior anomalies when implemented in their security and building control systems.”
These smart technologies work together to improve operations at factories, thus benefiting the customer by reducing workplace hazards, improving overall employee safety, and reducing energy consumption.
Top 6 video analytics features for factories
Jamie Barnfield, Senior Sales Director at IDIS Europe, pointed out that many of the key capabilities of video analytics are well matched to operational requirements in factory settings. The fact that video analytics have become more reliable and accurate, thanks to advances in AI deep learning, means that most industrial and manufacturing applications can now reap greater benefits from their camera systems.
1. Line cross detection
Line cross detection has long proved its value in security applications, with automatic detection of an intrusion at perimeters and vulnerable locations triggering alerts to events that busy monitoring teams might otherwise miss.
“Line cross is being put to equally good effect in manufacturing,” Barnfield said. “Virtual lines around potentially dangerous machinery restricted materials storage areas, and high-risk assets is improving safety as well as security.”
Forklifts and side-loaders can be detected when they enter restricted areas that could endanger other workers, enabling automated alerts, manual PA announcements, or configured to send managers and drivers warnings via SMS or email.
2. Fall detection
Fall detection is another tool that is increasingly useful in industrial settings, helping to ensure faster responses to what is a common cause of injury that requires rapid assessment and first aid to ensure the best possible outcome.
“This functionality allows easier and faster retrieval and export of footage when it comes to incident investigations to prevent fraudulent injury claims,” Barnfield added.
3. Fire safety
Video analytics can provide another layer of protection against fire risks in hazardous environments, with the ability to identify sparks or smoke before the fire has a chance to establish and spread.
4. Face mask and equipment detection
Face mask detection tools were honed to new levels of accuracy during the pandemic, with even the ability to detect when masks were not being worn correctly.
“The same tools are now helping to enforce hygiene protocols and eliminate human error in food processing plants and sterile manufacturing facilities,” Barnfield said.
5. License plate recognition
AI-powered LPR allows secure but smoother and faster entry to car parks and goods in/out, with live databases of pre-approved vehicles and drivers – and exclusion lists for vehicles known to present a risk - being used as part of increasingly sophisticated access and visitor management systems.
6. Smart search
Searching through thousands of hours of video footage to find incidents that are often only a few seconds long is one of the most difficult parts of an investigation. Fortunately, AI analytics can make this process easier.
“Adding value to all these tools is enhanced use of metadata which AI is enabling, making it easier and faster to search through video data to find people, vehicles, and events of interest,” Barnfield said. “Even where searches need to be conducted across multiple video streams, across extended periods, AI now enables searches within minutes, not hours or days.”
The benefits of video analytics in factories
Video analytics benefit factory customers in several ways. This includes better security and safety, improved compliance, operational efficiency, and optimizing costs.
Detect suspicious activity
Video analytics continue to strengthen security and improve risk mitigation, helping prevent losses and damage. Reliable automated detection of suspicious activity or events of interest means less chance of incidents being missed.
“This enables a more proactive approach to security and surveillance – intervention at an earlier stage - and it depends on the increased accuracy and reduced susceptibility to false alarms,” Barnfield pointed out. “This is what today’s deep-learning analytics are now delivering.”
Health and safety compliance
AI video in industrial and manufacturing settings is improving health & safety and underpinning compliance with legislation, including OSHA (in the US) and the Health & Safety at Work Act (in the UK).
Less staff, more efficiency
Another critical benefit – particularly important at a time of increased staff shortages and wage inflation – is the opportunity that AI provides to reduce staffing levels, take pressure off busy personnel in the control room, and monitor functions without negatively impacting operations.
“The benefits are cumulative, with officers freed from continual monitoring duties able do more rewarding work that adds greater value – for example, prioritizing first aid training and capabilities, being more visible and interacting more effectively with other employees,” Barnfield said. “This trend – which is helping to reduce security officer turnover - is further boosted by the new generation of phone and tablet enabled VMS apps which allow key video monitoring functions to be handled on the move.”
Production facilities can also utilize heat-tracking algorithms to recognize bottlenecks in manufacturing, process flows, packaging, warehousing, and shipping. The ability to view traffic, movement, and production patterns with heat maps provides facilities with the tools required to organize workflow, remove obstructions, and optimize layouts to increase operational efficiency.
Monitor staff movement
With most manufacturing still done wholly or partly by people, video analytics tools such as heat mapping can provide valuable insights into how staff moves around production lines. Analysis of recorded video data can reveal trends over time that might otherwise be less obvious – revealing movement inefficiencies and bottlenecks, for example.
“Factories and production lines can use simple motion detection, people counting, line crossing and other basic analytics features to increase safety, provide accountability, compliance and add situational awareness,” explains Roy Stephenson, Director of Business Development at Utah Yamas Controls. “For example, video analytics can be used to boost safety around machines and equipment by requiring the presence of operators at control and safety panel locations or by shutting down production lines when safety or security areas are broached.”
Ensure business continuity
Business continuity is a priority for many manufacturing operations, particularly those working to just-in-time principles, with even brief unscheduled interruptions to processes potentially leading to significant losses.
“Video analytics tools such as fall detection and line cross detection can help to prevent these interruptions by allowing faster and earlier interventions,” Barnfield said. “Whether it’s accidents, fire or floods, human error or malicious activity, where video analytics allows problems to be tackled early and prevented, losses can be reduced, and insurance premiums can be kept down too.”
Production lines in factories rely on a lot of electromechanical equipment, which is sometimes required to run 24/7. Unexpected failures of electrical or mechanical equipment can cause massive delays in production and costs companies thousands of dollars a minute.
“By utilizing thermal imaging cameras with heat detection analytics, production facilities can measure the heat of equipment such as pumps, switchgear, transformers, motors, Variable Frequency Drives (VFD), pulleys, belts other industrial equipment,” Stephenson said. “If temperatures rise outside of normal operating efficiencies, detecting it early can help operators to foresee potential equipment failures which could result in downtime, interruptions, or production jams.”
Quality control and defect detection
Production facilities can utilize fully automated machine vision cameras for repetitive tasks such as object counting or defect detection by using advanced analytics to verify quality control which in turn enhances productivity and accuracy.
“A camera on a production line can compare each product to a predefined image to look for any imperfections. This can increase the speed and accuracy of production lines and reduce operational costs,” Stephenson added.