Today, our industry and the greater technology sector are only at the beginnings of understanding how to truly harness the possibilities of metadata. To do this, we are simultaneously pushing the boundaries at different directions – upwards to explore and innovate new ways of mining intelligence from metadata, and outwards to make metadata ­collection and analysis easier and more accessible at a mainstream, mid-market level.

Standardization is key to making these technologies accessible, creating common pathways, definitions and approaches that are available to all.

Onvif has already made significant strides in this area with Profile M, created to standardize the handling of analytics metadata between cameras, VMS and other software platforms. This addresses metadata at a foundational level, while at the same time opening access to IoT environments and enabling cloud-based solutions and services for manufacturers, systems integrators and end users and other organizations looking to add value using metadata. Standards such as Profile M help to enable greater innovation in this area and the ability to scale the solutions that are developed to the larger market.

Onvif membership boasts some of the leading IP video companies involved in this research and development. Here are some of their thoughts on the evolution of metadata and its future impacts.

The Evolution of Metadata and Future Impact

Gunnar Erlandsson, Rudolf Spielberger, Barry Norton and Todd Dunning answer GIT SECURITY’s questions on metadata and the future market.

GIT SECURITY: How is the security industry using (video) metadata today?
Gunnar Erlandsson, Global Product Manager, Analytics, Axis Communications: Metadata is becoming increasingly important. Automatically generated metadata with information about what’s happening in a scene makes it easier to build tools to help operators handle large amounts of video data. Metadata can be used in the video management software context, to enable automated or human responses to alerts, and to dramatically decrease the time needed for video searching. Another integration possibility is to collect the metadata as statistics in dashboards, for example visualizing summaries and trends, that can give actionable insights.

Rudolf Spielberger, Solution Business Partner Program, Bosch Security and Safety Systems: Video analytics, a form of Artificial Intelligence, has long produced metadata to add sense and structure to recorded video footage, which helps users to understand scenes and situations. Metadata makes it possible to retrieve the correct footage from hours of stored video. It also analyzes scenes to identify potential undesirable events and provides various statistics so users can act faster and more efficiently or make smarter business decisions. It’s the main reason Bosch video security cameras have had built-in video analytics as standard since 2016.

Barry Norton, VP of Research, Milestone Systems: The security industry and Milestone see a significant increase in the use of structured data, including video metadata and events. At Milestone, we are continuously building capabilities of the ingestion and utilization of such data.

Todd Dunning, Director, Product Management, Video Security & Analytics, Pelco: Pelco by Motorola Solutions used metadata in analog video solutions 20 years ago with Pelco D protocol sending date and time metadata from the camera to the VMS. As the industry migrated to digital video, the ability to expand metadata and standardize methods for metadata has increased. Today, Onvif Profile M provides a standardized method to provide metadata for geolocation, vehicles, license plates, and human faces and bodies.
 

How do you see the use of metadata changing in 2025 and beyond?
Todd Dunning: Metadata will continue to expand into operational areas of camera technology. Temperature ranges (thermography) in thermal technology, crowd behavior reports, camera health metrics, and helmet and safety vest metrics are a few items that customers are asking for and are seeking simple ways to report on through metadata.

Barry Norton: Over the next 3–7 years, the data on which security solutions are built will become both richer and more domain specific. This will be driven by increased focus on verticalization, the growing sophistication of cameras and analytics producing metadata, and by the cross-fertilization with other devices and data sources.

Rudolf Spielberger: More powerful hardware and significant metadata growth will drive increased adoption of deep learning-based video analytics. Deep learning offers more in-depth scene understanding and enables informed decision-making for quick and automated responses. For example, our Bosch traffic monitoring solution, founded on deep learning, produces comprehensive and reliable metadata. It achieves accuracy levels beyond 95 percent for real-time event detection and data aggregation. The technology will only improve, allowing us to create more value for our customers.

Gunnar Erlandsson: We expect metadata to become more important and widely deployed. With the advances in AI technology, it will become possible to generate even richer metadata, including more information about what’s in the scene.

How important are standards to the evolution of metadata?
Rudolf Spielberger: Adhering to common standards – such as Onvif Profile M for metadata and events for analytics applications – is vital to making metadata reusable and ensuring system interoperability. Combining different metadata sources for verification purposes or building solutions that work with metadata from other sources will help our customers create even more thoughtful and tailored solutions, far beyond security and safety alone.

Gunnar Erlandsson: Standards are very important. Onvif Profile M has been instrumental in the security industry’s adoption of metadata to ensure effective interoperability between system components. Further, customers can trust that they do not rely solely on one vendor and do not risk being stuck with legacy systems that cannot be expanded.

Todd Dunning: Standards make integration easier, whether it is across companies or within a company.  Documented ways to communicate often accelerate end solutions for customers. Although standards are not always the fastest path forward, in the end, it helps manufacturers anticipate technical requirements, and can reduce costs and time for integrators granting them brand flexibility in finding solutions for end users.

Barry Norton: Standards are essential in moving beyond basic and generic video analytics to more complex domain-specific solutions. Standards lower the cost barriers and therefore accelerate the adoption process because they scale beyond a single vendor and allow a common interpretation of interconnected data, including video metadata, across the value chain.

This is an article of our jubilee issue. Click on the button to find all of the interviews, articles and the e-paper.