Perimeter protection refers to the security measures and systems implemented to safeguard the boundaries or perimeters of a physical location or property. The primary objective of perimeter protection is to prevent unauthorized access, detect potential threats, and deter intruders from entering protected areas.
Traditionally, the analysis of live video streams used to be manual, relying on human perception for visual identification of events happening in each feed. Nowadays, deep learning at the edge is enabling the automation of the analytics task, thereby allowing for easier scalability, and improvement in overall performance. Having a Perimeter Protection system on the edge enables real-time and immediate response to security threats, as the AI algorithms can process and analyze large amounts of data locally at the edge, without relying on centralized processing and the robustness of network connectivity. This reduces latency and ensures faster detection and notification of potential breaches. Additionally, by deploying AI on the edge, the system can continue to operate even in the event of a network outage or disruption, maintaining the security of the perimeter, or in remote and disconnected locations such as military bases or underground facilities. Edge-based AI also minimizes the amount of data that needs to be transmitted to the cloud, enhancing privacy, and reducing bandwidth requirements and costs.
The Hailo AI processors are transformational to these applications as they allow faster and more accurate event identification at a lower cost.
Hailo offers significant cost-savings on video analytics with more compute per price unit, that is translated into more complexity per application, more applications per stream and more streams per platform. It also enables significant saving on streaming bandwidth and storage space thanks to event-based recording
Hailo’s AI processors are designed to withstand harsh conditions and conform with industrial operation conditions, applicable for perimeter control systems installed in outdoor and sometimes challenging environments
Perimeter control systems require processing of people’s personal information. Video analytics on the edge means that only video metadata, and not personally identifiable information (PII) needs to be transmitted and stored in the cloud
Multi-camera and multi-person re-identification (Re-ID) refers to the task of identifying individuals across different cameras in a surveillance system, or in several occurrences over time through a specific camera, for security or statistical analysis purposes. The task includes the identification of a specific person multiple times, either in a specific location over time or along a path between multiple locations. It involves matching the appearance of a person captured on one camera with the appearance of the same person captured on a different camera, or on the same camera at a different times while accounting for changes in lighting, pose, and occlusion.
AI enables multi-camera Re-ID as this task requires processing and analyzing large amounts of visual data from multiple cameras simultaneously, which is challenging to the level of impossible for a human operator. AI models can learn to extract and represent meaningful features and personal attributes from images or video frames, such as body shape, clothing, and accessories, and use them to match and identify individuals across different video footage. Re-ID on the network edge means that people can be identified and re-identified with improved privacy protection, as only anonymized information will be streamed and stored in the cloud.
Video metadata refers to the descriptive information associated with a video, which includes various attributes such as person identification, face attributes, and other relevant characteristics. These attributes can be extracted and analyzed using AI to derive valuable insights and enhance various applications.
AI is used in video metadata for multiple applications, including:
The better, and more elaborate the video metadata is, the better the results and insights will be. The quality of the metadata relies on the quality of the analytics, and this is where Hailo’s advantage in enabling to run advanced algorithms at the edge come into play.
