The JetStream is a multi-channel video AI carrier board for Jetson modules. The JetStream provides dual HDMI and eight composite video inputs, allowing multi-channel video AI applications to be rapidly developed and deployed. This SWaP optimised solution is ideal for demanding applications in military, communications, transportation, mining and energy industries.
The JetStream features dual HDMI/DVI inputs at up to 1080p and an 8-channel NTSC/PAL/RS-170 low latency video capture engine. All inputs can be simultaneously captured for processing by the Jetson Xavier NX module, allowing vision-enabled AI applications to use a wide range of standard video sources. An optional MIPI expansion module provides expansion to other video interface types on request.
The JetStream has a HDMI output for local display and Gigabit Ethernet for connectivity. An M.2 (Key M) slot is provided to allow connection of high-performance NVMe storage. Additional expansion is possible via the PCIe/104 OneBank bus connector.
The JetStream is a standard PCIe/104 form factor card and is powered from 9 – 15 V DC inputs. The JetStream is supported by a comprehensive board support package and a series of application notes designed to help rapid development of vision-enabled AI applications.
Applications
• AI inference - object recognition, ANPR, pose detection
• Device edge computing
• Unmanned vehicles (UAV, ROV)
• All-round real-time situational awareness
• Traffic monitoring and control
• Video acquisition and analytics
• Remote video surveillance
• Rugged intelligent video recorders for marine, aviation
• Border security
Features
• Support for NVIDIA Jetson Xavier NX and Nano modules
• 2 x HDMI video inputs at up to 1080p
• 8 x composite PAL/NTSC/RS-170 video inputs
• HDMI 2.0 display at up to 4K resolution
• Multi-channel AI video inference
• H.264/H.265 video encoding and streaming
• Rugged, real-time video capture/analyse/stream solution
• M.2 (Key M) NVMe storage
• Gigabit Ethernet
• PCIe/104 OneBank expansion
• Industry standard PC/104 mechanical form factor
© Technews Publishing (Pty) Ltd. | All Rights Reserved.