New Uses for AI
- By Aaron Saks
- Jul 22, 2024
New applications of AI in IP cameras are delivering precise detection, robust search capabilities, elevated data analysis and enhanced image quality. When combined with built-in analytics, AI can help streamline forensic investigations and also supports several mission-critical business functions.
Boosting Performance
Many companies, Hanwha included, first started introducing AI into their products on a selective and specific basis. Now, AI is a key feature across most product lines. Many new and existing products now incorporate AI technology to boost their performance to unprecedented levels and enabling new vision solutions that address customers’ complex challenges by adding new layers of business intelligence.
This trend is about more than devices. It is all part of the continuing convergence of hardware devices and software solutions -- combining 24/7 protection with the latest advancements in AI, analytics, and cloud-based management to create data-driven and analytics-based platforms.
New types of business intelligence software can harness the data gathered by the embedded edge AI analytics in IP cameras to monitor market trends and events in real time. These software applications can process metadata and presents contextualized data through customizable widgets and charts in a visualized dashboard. This gives customers context about their facility and operations, deriving insights that can turn unrealized data into actionable insights.
New Types of Surveillance Applications
AI is enabling a range of surveillance applications, such as remote system monitoring, cloud-based services, analytics, and data gathering. AI allows cloud and remote monitoring solutions to effectively manage bandwidth by only alerting teams when an important event is triggered, such as a person in an area, loitering, etc.
Previously, cloud recording was limited to continuous recording and processing motion detection in the cloud, whereas now edge AI processing can be coupled with a cloud solution. Many remote monitoring solutions would use simple SMTP email notifications and short video clips.
Now, detailed metadata can be transmitted allowing an operator to quickly see relevant information, while only receiving notifications on items that are pertinent. When an incident occurs, locating a person of interest can take a matter of minutes instead of having to spend hours sifting through hundreds of camera streams.
AI cameras are also being used to improve operational efficiency in “non-security” applications. A common example is workplace safety. If there are areas of a warehouse or manufacturing facility, for example, where people should not normally be present, AI can provide real-time altering and recording of times when this rule is breached.
In retail stores, AI-powered cameras can fit marketing and merchandising opportunities, looking at heatmaps to see hot/cool areas of the sales floor, traffic patterns, etc. This allows the retailer to adjust endcaps and other product placement without needing another siloed system.
AI Accuracy
To ensure the highest level of accuracy possible in AI cameras and devices, it is important to think about how you want to use the cameras and ensure that they have the proper field of view, which impacts the pixel density and AI performance. Depending on how you want the camera to function, ensure you have the right resolution, zoom, and angle when spec’ing and installing the camera. In addition, make sure the image is optimized, thinking about WDR and exposure settings and IR/lighting.
The continued integration of AI cameras with other security systems can create more proactive and intelligent total security and surveillance solutions. The goal is to make alerts more meaningful by removing false positives from everyday nuisances, such as lighting changes, shadows, small animals, etc.
Furthermore, we want to remove data from legacy silos so that different systems can make use of this data. A common example is visual verification of access control alarm systems. Using AI, a rule could be written that if a person is not seen at a certain doorway, to ignore the alert. Or an operator verifying the alert can quickly look for people in the forensic search instead of having to manually review lots of footage. They can then search for those attributes (clothing color, etc.) to see if there have been repeated attempts.
Looking ahead, we will continue to see emerging technologies being integrated with AI to create more comprehensive intelligent monitoring and surveillance environments. For example, cameras or audio sensors are being AI-enabled to make them more accurately detect certain scenarios based on the monitored audio alone, while ignoring false positives. Thermal and bi-spectrum cameras are gaining AI capabilities allowing them to provide thermal imaging with AI detections. In the future, we will see this continue to devices such as RADAR or LIDAR sensors and beyond.
AI models will also become more specialized, allowing customers to choose what they want to detect, allowing for vertical specialization without needing specialized cameras, such as detecting shopping carts and forklifts. This diversity across sectors shows just how far-reaching AI has become, not only for surveillance but for every market segment.
This article originally appeared in the July / August 2024 issue of Security Today.