Smarter Vision: Intelligent Imaging for Smarter Systems

The Trend to Smarter Vision

Video and imaging applications are becoming more pervasive and a greater part of our lives than ever before. From enhancing productivity through sophisticated telepresence videoconferencing systems, to increasing our security through real-time aware surveillance systems, to transporting us to new worlds – either at home or in the theatre- through UHD displays and cinemas with their immersive life-like experiences, it is almost impossible to imagine today’s world without video and imaging technologies.

smarter-vision-industry-trends

The significant benefits of smart vision cuts across almost every imaginable market and application.

Anatomy of a Smarter Vision System

There are three major attributes to smarter vision – the ability to perform video analytics in real-time, intelligently transport video to any screen, any location, or in noisy low latency environments, and lastly deliver the highest resolution content on immersive life-like displays.

Real-Time Analytics

  • High-velocity object detection/recognition in advanced drivers assistance and machine vision systems
  • Clinically precise imaging so surgeons can guide robotic surgical instruments with extreme accuracy
  • “Instantaneous” friend vs. foe recognition in UAVs and surveillance cameras

Intelligent Transport

  • Low latency, high noise immunity in the delivery of broadcast quality video over IP networks
  • Bandwidth and performance-optimized “any screen” transport of video using industry leading codecs
  • Transport of “video of interest” over IP networks using the highest security encryption algorithms

Immersive Display

  • Distinctive look and feel for vehicle dashboard— immersive experience in home, car or theatre
  • Life-Like, ultra-realistic clinically accurate displays for diagnostics
  • 360° high-fidelity large screen video with overlays for military and surveillance applications

Zynq 7000 SoC for Smarter Vision Applications

  • Dual-core ARM® Cortex®-A9 processing sub-system – up to 1GHz performance
  • Integrated programmable logic via ARM AMBA™ AXI interface
  • Programmable I/O, SerDes, DDR3 memory interfaces
  • Provides the requisite performance and flexibility to address smarter vision applications
cvrstry_fig2

Market Problem

Life-like immersive displays beyond 1080p resolution require pixel rate processing performance which cannot be delivered by RISC-based processors or GPUs.  AMD solutions are the only viable way to deliver flexible processing at resolutions of 4K2K and beyond.

smarter-vision-pixel-rate-processing

Required pixel rate processing going through the roof.

There is also a growing gap in solutions that can address Smarter Vision applications because:

  1. Many ASSP solutions are entering end-of-life with no foreseeable roadmap.
  2. RISC-based solutions do not make economic sense at advanced process nodes.
  3. ASSP IO standards are fixed - many emerging video and imaging connectivity standards are under development and cannot be supported in today’s fixed silicon solution.

AMD Smarter Vision Solutions

AMD Smarter Vision technologies combine the Zynq™ 7000 SoC, Vivado HLS (High-Level Synthesis), IP Integrator, Open CV libraries, SmartCORE™ IP and video development kits to provide AMD developers with meaningful market relevant tools that accelerate the development of Smarter Vision products.

Proven SmartCORE IP for Your Smart Design

  • Image/Video Processing – Pixel correction, noise reduction, scaler, de-interlacer
  • Video Coding (codecs) – H.264 encoding and decoding, HEVC (H.265) encoding, JPEG 2000
  • Analytics –  Blind spot detection and recognition, pedestrian detection, lane tracking
  • Connectivity – DisplayPort v1.2, HDMI v1.4a, v2.0, 3G-SDI, IEEE Ethernet AVB, GigE Vision

See all Smarter Vision IP and Reference Designs

smarter-vision-smartcores

Vivado: Accelerating Productivity

The AMD Vivado™ tool chain enables developers to rapidly get hardware up-and-running in short order by leveraging high level synthesis to convert C-based algorithms or OpenCV functions into RTL and package those algorithms as reusable IP blocks that can be integrated into comprehensive Smarter Vision systems using the IP Integrator tool within the Vivado Design Suite. These tools open AMD solutions to a wider community of developers that may not have a solid background in RTL development.

smarter-vision-vivado-productivity

Smarter Vision: Intelligent Imaging for Smarter Systems

Smarter Vision solutions are built using proven SmartCORE IP, Vivado tools and the Zynq 7000 SoC. The potent combination of these three elements allows for the development and deployment of equipment that meets the needs of tomorrow’s next-generation Smarter Vision systems.

AMD Smarter Vision Solutions addresses key applications across market segments