Specifications
Models
FantoVision40-CXP12 (CoaXPress 12), FantoVision20-GigE (10GigE Vision), FantoVision20-CL (Camera Link)
AI Compute
NVIDIA Jetson Orin NX — up to 157 TOPS
FPGA
Altera Arria 10 FPGA for frame grabbing, ISP and pre-processing
Acquisition Bandwidth
Up to 50 Gb/s deterministic real-time ingest
CoaXPress
Quad CXP-12 links with up to 13 W per cable (PoCXP)
GigE Vision
10 GigE Vision, up to 20 Gb/s total bandwidth
Camera Link
Camera Link Deca configuration
Vision Throughput
Over 1 gigapixel per second acquisition, processing and compression
Offload
FPGA handles ISP and pre-processing, cutting host CPU load
Compression
Real-time JPEG compression with adjustable ratios to cut bandwidth and storage
Architecture
Jetson (AI) + FPGA (I/O and pre-processing) heterogeneous design
Dimensions
134 x 90 x 60 mm
Weight
750 g
Ruggedisation
Industrial-grade ruggedisation for harsh environments
Target Platforms
Payload-sensitive drones and robotics where SWaP-C matters
Recognitions
Gidel SkyBoost-RT — Vision Systems Design 2026 Innovators Awards Silver Honoree
Software
Gidel image-processing IP with NVIDIA TensorRT integration
Overview
Gidel's FantoVision family is an edge AI vision computer that pairs an NVIDIA Jetson Orin NX module with an Altera Arria 10 FPGA. The design deliberately splits work: the Jetson runs the neural network while the FPGA performs frame grabbing, image-signal processing and pre-processing, which is described as breaking the I/O wall of a standard Jetson system limited to USB, MIPI and standard GigE.
Three interface-specific models target the acquisition standards high-speed inspection actually uses — FantoVision40-CXP12 for CoaXPress 12 with quad CXP-12 links and up to 13 W per cable of power over CoaXPress, FantoVision20-GigE for 10GigE Vision with up to 20 Gb/s total bandwidth, and FantoVision20-CL for Camera Link Deca. The systems sustain deterministic real-time bandwidth of up to 50 Gb/s and can acquire, process and compress more than a gigapixel per second, with adjustable JPEG ratios to reduce storage load.
Physically the units are 134 x 90 x 60 mm and 750 g, an SWaP-C optimisation that puts workstation-class I/O into a palm-sized chassis suitable for drones and mobile robots, with industrial-grade ruggedisation for harsh environments.
Key Benefits
Jetson Orin NX at 157 TOPS plus FPGA pre-processing. CoaXPress 12, 10GigE Vision and Camera Link in dedicated models. 50 Gb/s deterministic ingest and 1 gigapixel/s throughput. 750 g, 134 mm chassis for SWaP-constrained platforms.
Applications
High-speed inline inspection, multi-camera industrial vision, drone and mobile-robot perception, semiconductor wafer inspection, sports and broadcast analysis, and any CoaXPress or Camera Link pipeline that needs on-board AI.
Request a Quote — GIDEL FANTOVISION — JETSON ORIN NX + FPGA EDGE AI VISION SYSTEMS
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →