Page 16 - Security Today, November/December 2022

Computer vision, essentially, is a set of eyes that can analyze what is happening within its field of vision. Because today’s cameras have perception capabilities that far exceed the human eye, com- puter vision can be used to implement and monitor safety and quality assurance protocols more accurately and effectively. A bottling company no longer needs to rely on a human to notice a half-filled bottle—a camera armed with computer vision will see it every time. Likewise, a camera equipped with thermal monitoring capabilities can track the temperature of machinery and alert if it becomes too hot or too cold. Here, again, DLPUs are essential—andd ubiquity is making computer vision technology commonplace. In the past, an orga- nization would need expensive and highly-specialized cameras in order to use computer vision. Today, DLPUs are not just for high-end cameras—they are becoming the standard. That means just about anyone with a truly modern surveillance camera can leverage computer vision for a wide range of uses. Better still, it reduces the number of devices organizations need to deploy, as the same camera they already use for security monitoring is now used for quality assur- ance and other purposes. There are also devices today that embed the DLPU functionality into their own chip which can minimize device size and mitigate heat issues at the same time. Innovation continues at accelerated speed. Trend 4. Interoperability Will Make Organizations More Efficient As computer vision allows companies to use individual devices for multiple purposes, organizations will be able to allocate re- sources more efficiently. Security teams, facilities teams, market- ing teams, and others can all use the same camera deployments for wide-ranging purposes. This means that an organization can reduce the number of devices in use, and it can draw funding for those devices from across multiple budgets. This might not seem VIDEO METADATA: DESCRIBING THE DETAILS THAT MATTER Modern surveillance systems generate an overwhelming (and mostly unused) amount of data. This is especially true when record- ing video in 24/7 operations, which is es- sential to capturing evidence, incidents and events. It is not only hard to pick out what really matters in a scene, but also extremely time consuming. Making data more iden- tifiable and actionable is a key problem to solve. Applying metadata to describe key details in a scene allows data to be more identifiable and actionable. This is why metadata is the foundation for gathering intelligence from surveillance video and/or audio streams. Metadata pro- vides a fast way to find, evaluate, and act on the singular details that matter the most through one, hundreds or thousands of video and audio footage streams. Metadata is now an essential part of effective security and business operations. WHAT IS METADATA? Typically, Metadata is referred to as ‘data about other data.’ In the context of video surveillance, that translates to ‘data about video data’. Video metadata accurately de- scribes the details that matter in a scene. For instance, attributes for metadata can describe all sort of details about moving ob- jects of interest, e.g. location, time, colors, sizes, shapes, coordinates, volume decibels, speed, direction, etc. Additionally, foundational details can be added, such as video stream description, co- dec, time stamps and device identity. The aforementioned are ‘meta’ descrip- tions of details in, or related to, a scene. Based on AI machine and deep learning, Meta descriptions can be more (or less) granular. This allows for classifying a group of pixels as a person, animal, vehicle or oth- er pre-defined object classes. Being more precise with more refined descriptions of people or objects e.g. vehicle type, make model, color, speed, direction, etc. THE VALUE OF METADATA Metadata not only provides details about people, objects and events in a scene. It also allows large amounts of video and recorded footage to quickly group, sort, search, re- cover and use. As a result, the overall use cases for metadata fit into three areas. 1. Real-time alarm triggering and notifica- tions 2. Post event forensic searching 3. Statistical analysis and reporting ADDING INTELLIGENCE TO SCENES Metadata essentially assigns digital meaning to each video frame about the objects and events within it. In other words, it adds in- terpretation or intelligence about the scene rather than just the raw video footage, which needs to be processed manually by an operator. Once software can interpret scenes in this way, it can understand the scene details and enable the scene to be acted upon in re- al-time via events, after events (post-event), via manual search or simply analyzed for statistical analysis. This enables the use of metadata to design baselines that define what is ‘normal’ for any scene feed from any individual camera. In turn, this allows soft- ware to recognize any degree of deviation, anomaly or specific behavior or activities, etc. as well as predict what will happen in that scene to a specific probability. HARNESSING THE FULL POTENTIAL OF METADATA Video metadata adds immense value to a video management system. In fact, its true potential is realized when applied to mul- tiple inputs spanning visual, audio, activity, and process-related inputs. In the manage- ment of any site, things like RFID tracking, GPS coordinates, tampering alerts, noise de- tection, and point of sale transactional data, are all high value data sources. Unifying this metadata generated from many different sources means gaining much more insights than one can ever get from each system alone. Interoperability is key, and open-pro- tocols and industry standards are essential to this effort. Ultimately seamless metadata integration will allow us to harness massive amounts of data from all kids of systems and gain a greater understanding of everything around us. Joe Danielson, Global Enterprise Solutions, Axis Communications, AB. 16 NOVEMBER/DECEMBER 2022 | SECURITY TODAY COVER STORY