NEXVISION is a hardware and software design house specialising in vision systems and video over IP network technologies, and they are developing a 4K camera.
You may remember Nexvision as the makers of the V.I.O. POV.HD Professional POV camera.
Here are the few details coming from the R&D devision of Nexvision on the Maestro 4K camera.
The Maestro is a 4K video camera project based on the Nexvision Cam Master + reference design.
Designed at the state of art, this video camera will embed a high-end optronic system :
Nvidia Tegra X1 processor
Xilinx Ultrascale FPGA
Sensor CMOSIS CMV12000
Our latest Image processing algorithms from our laboratories
NVIDIA TEGRA X1 PROCESSOR + XILINX ULTRASCALE FPGA:
Overview:
Nexvision’s CAMMASTER smart camera is a complete movie camera reference design
Designed for 4.8 GigaPixel/s processing capability
On-board video analytics (Nvidia Tegra X1)
Embedded Linux dedicated distribution
Highlights:
Multiple extra high resolution independent video inputs from 2 (Full-HD) to 20 (X-HD™) megapixels
H264-MP codec : 4K 30fps
On-board dedicated video enhancement image pipe (HDR, FPN, Noise reduction, EDoF), Nexvision’s IP : PIXIP™
GigaEthernet, USB2.0, USB3.0, SATA, CoaXPress, 6G-SDI, Audio, HDMI video output, PCIe, SPI, I2C, RS232, NVMe, HMC
Video analysis framework (third party accessible)
Onboard video recording and meta data storage (SSD, SD, NVMe)
Onboard streaming server based on our NexStream™
Features:
Extra high resolution from 2 megapixels (Full-HD) to 20 megapixels (X-HD™)
Full framerate, high quality video encoding
On-board high-end FPGA with three DDR4 SDRAM banks, dedicated to video enhancement and advanced video processing (NexVision’s IP : « PIXIP »)
2×300 megapixels/s MSOC’s video processing pipeline, two dedicated DDR4-SDRAM bank Codec (Tegra X1)
H264-Main Profil for minimal storage size and bandwidth : 120M pixel/s – Full HD 1080P (2 M pixels @ 60fps)
Optional video codec : Motion-JPEG, MPEG4-part2, RGB444 , YUV422, RAW, Cinema DNG
Image pre-processing (FPGA)
Media processor image processing pipe
Advanced 12-bit based color processing for high fidelity color reproduction (3D LUT)
Temporal noise filtering and contrast enhancement
Multiple exposure blending provides realtime HDR for high details retention in low and over exposed area
Super resolution
Multisensor capable
Multispectral band image sensor fusion (Visible, Short Wave IR, Thermal IR)
Multisensor panoramic or 360° immersive video realtime on board unwarp
Audio/Video Streaming
Live video streaming
Low latency path for real time viewfinder
Full Duplex Audio
Open onboard video analysis, for example : Video analysis framework, third
Image Sensor Board Compatibility
Video interface is compatible with any sensor ;
Progressive scan, color with 2, 5, 8, 11, 12, 17 or 20 megapixels (X-HD™)
High sensitivity up to 12V/Lux-sec, low read noise <2e-
Selectable video resolution and frame rate by image sensor daughter boards choise Up to 16 sensor boards, any combinations of Nexvision’s board :
“ISB17M” board : 17Mpixel, 5μm pixels
“CAM12M” board : 12Mpixel, 5.5μm pixels (22.5×16.9mm), 4.64V/lux.s, 12Mpixel 4096 x 3072 pixels 300fps shoot & preprocess frame rate, H264@60fps
“CAM2M” boards : FullHD 1080P(1920×1080) 340fps shoot & preprocess frame rate
“CAMPANO” boards : 4096 x 500 multisensor board, very high sensitivity 12V/lux.s
“SWIRCAM” : Short Wave Infra Red, 320×256, 100fps, with TEC cooling controller
Video Input / Output
RAW video interface on FPGA – 100 x LVDS Pairs – 320 pins connector
GENLOCK IN or OUT + Composite video output – MCX connector
1 x 3G-SDI IN or OUT by software – MCX connector
4 x CoaxPress CXP6 IN or OUT by hardware assembly – MCX connector (Power over Coax output with 12W@24V)
Linear Time Code (LTC) IN and OUT – Lemo5 connector
LCD display (with OPENGL) – HDMI 2.0 (with audio) – HDMI 19 pins connector
D1 PAL/NTSC composite analog output from FPGA – MCX connector
4 x 6G-SDI OUT (can be used as SD, HD, or 3G-SDI) – MCX connector
For more: Nexvision, and Cam Master+ details.