Custom camera platforms
We have made a name on building extraordinary camera systems for drones, robots, medical devices, autonomous cars etc., thanks to our deep expertise of all things related to machine vision: from working with pre-release edge computing FPGA and GPGPU modules to developing customizable BSPs, drivers and edge AI applications. See how our platforms can be used in your project.
Powerful edge computing
Antmicro works with bleeding edge image processing technologies, employing the latest, smartest and most powerful edge computing platforms. Thanks to our close partnerships, we are often among the very first to have access to pre-release silicon and hardware. Coupled with our early open source board designs, we are happy to equip our customers with knowledge and market advantage to quickly proceed with more sophisticated commercial projects.
Expandable vision platforms
Dedicated carrier board for NVIDIA’s most powerful edge platform, Jetson Xavier AGX. Features fully integrated Allied Vision’s modular Alvium MIPI CSI-2 cameras. Possible to interface with up to 8 cameras, intended for high-end edge AI applications. Customization, software and AI development services available.
Industrial 3D stereovision camera system
MOREA high-speed stereovision camera system for industrial use cases. Hybrid FPGA+GPGPU edge computing platform with Xilinx UltraScale+ and NVIDIA Jetson TX2 for AI-supported 3D vision and tracking in real-time. Supports a dedicated stereo camera module with liquid lenses in an industrial-grade casing. The modular and scalable design allows for complex data fusion and integration with larger industrial systems.
Officially co-released with NVIDIA on the day of the Jetson Nano premiere. Very small footprint Jetson carrier board compatible with the Jetson Nano (both devkit & production pinout) and the compact Jetson Xavier NX. Enables up to 4 MIPI CSI-2 cameras. Supports Allied Vision cameras and Antmicro’s camera accessories. Fully open source hardware, released on GitHub. Offered as part of Antmicro’s product development services.
Open Zynq Video Board
MOREA video-oriented carrier board for Xilinx Zynq or FPGA modules with HDMI, Ethernet, and SD card support. Features a pair of USB 2.0 host ports for external peripherals and a set of PMOD connectors allowing to connect external sensors and actuators. The board allows interfacing up to two 2-lane MIPI CSI-2 cameras. Featuring hard dual-core ARM Cortex-A9 SoC or soft RISC-V SoC running Zephyr/Linux. Both hard and soft processing system can be simulated in Renode/Verilator. Open source HDMI and MIPI CSI-2 IP. Full open source hardware, released on GitHub.
Apalis Smart Vision Baseboard
MOREExpandable, multi-purpose Apalis iMX8 or TK1 SoM hardware platform with dual-camera board, ideal for gaming screens, smart devices, on-board computers, rich UI, video processing: AI-capable, up to 3 MIPI CSI-2 cameras, USB 3.0 and mini-PCIe. Works with Toradex Linux BSP. Extra AI capabilities can be enabled. Customization and Android services available.
TX2 Deep Learning Platform
MOREPopular, Linux-enabled hardware platform with Antmicro’s feature-rich carrier board for the NVIDIA Jetson TX2 bundled with three dedicated dual camera boards. Allows for up to 6 MIPI CSI-2 cameras, e.g. for 360 degree vision. Available as a stand-alone baseboard as well as the basis for more complex product, software and edge AI development services.
Open Source Coral baseboard
MOREThe Coral Baseboard is an open-source carrier board designed by Antmicro for the Coral SoM, a low-cost AI platform featuring Google’s Edge TPU ML inference ASIC and NXP’s i.MX8X SoC. A great fit for embedded machine learning applications that need to execute state-of-the-art vision models in a power-efficient manner, the baseboard can be interfaced with our camera boards and accessories and be customized by Antmicro for specific AI projects. Coral's availability in other variants such as standalone chip and PCIe/USB modules make it a compelling, scalable choice for intelligent devices we help our customers build.