Video Analytics at the Edge

Enabling-Generation Distributed Video Analytics

Today's sensing, processing, storage and connectivity technologies, available in high density, low power packages enable the next step in distributed video analytics, for each camera self is a server. The new IP cameras and camera technologies coupled with micro server like compute, storage and I/O resources plus sophisticated software are able to form highly scalable edge analytics for surveillance and security, parking, toll management, and other real-time analytics where the monitoring takes place.

Video analytics at the edge minimizes video processing and the backend, which simplifies balancing the entire system from the edge to the datacenter. By doing the analytics at the edge with capable hardware such as good quality MIPI-CSI sensors and software, companies can acquire higher quality, uncompressed video streams at the camera without detrimental compression artifact that can creep in downstream. Companies have the option of passing only the results as meta data, saving potentially expensive bandwidth and accelerating the analytics and data searches in the backend.

Video Analytics at the Edge CuttingEdgeVideo

This edge based approach, with high-capacity SSD's for fast, local storage allows uninterrupted monitoring should the connection to the data center go down. A sophisticated video analysis algorithm means instant actionable intelligence concerning your protection objectives as readily available.

From the edge to the cloud - with video analytics at the edge, cloud customers can implement different cases with local image processing that sends only metadata of video streams back to the cloud data center.

Powerful media processing and analytics support - the Intel AtomTM processor E3800 series with four cores and supporting hardware delivers the compute and I/O capabilities that video analytics at the edge demand.

Strong protection for platforms and video data - intelligent IP camera heads connected to the network can be vulnerable to attack without adequate security support the same Intel Atom processor integrates Intel Advanced Encryption Standard New Instructions enabling very fast processing of today's advanced algorithms. Many intelligent IP camera heads are designed with remote manageability.

Bridge support for quick innovative solution designs- the video design community has rich support for innovative designs some of the following resources are Kraftway, Kedacom, TheopensourceMediapipeFrame, Open source Linux, LibXcam framework.

Using existing frameworks, available open source software, and rich video hardware and software available in the ecosystem, designers have access to repeatable building blocks for fast time to market of innovative video analytics solutions at the edge of the network.