Product Brief
Overview
Block Diagram
The diagram below illustrates a design based on the Ambarella CV2S device.
Key Features
CV2S - Computer Vision SoC
for IP Cameras
The CV2S chip targets IP camera designs
Flexible Low-Power Platform
• 10 nm low-power CMOS process
• 64-bit quad-core Arm
®
Cortex
®
-A53 CPU
up to 1.0 GHz
• Linux kernel version 4.14+ or newer (64-bit)
•
®
vector processor for CNN/DNN
algorithms
• OTP, secure boot, TrustZone
, IO
virtualization
• Industry leading image sensors support
4K Computer Vision Engine
• CNN- / DNN-based processing: detection,
classification, tracking, and more
• CNN performance up to 20 times of CV1
• Dedicated CVflow computer vision processor
• Tools for high- and low-level algorithm
development
• Dedicated hardware for stereo processing
• CNN toolkit for easy porting with Caffe,
TensorFlow, and ONNX
• Open SDK
Advanced Image Processing
• More than 800 MPixels input rate
• Multi-exposure line-interleaved HDR
• Hardware dewarping engine support
• Electronic image stabilization (EIS)
• Six independent sensor inputs
• 3D motion-compensated temporal filtering
(MCTF)
• Superior low light processing
High-Efficiency Video Encoding
• H.265 and H.264 video compression
• Flexible multi-streaming capability
• Up to 4Kp30+1080p30 video performance
• Multiple CBR and VBR bit rate control modes
• Smart H.264 and H.265 encoder algorithms
Ambarella’s CV2S SoC combines image processing, 4Kp30+ video encoding,
and CVflow
®
computer vision processing in a single, low-power design.
The CV2S’s CVflow architecture provides the deep neural network (DNN)
processing required for the next generation of intelligent IP cameras.
Fabricated in advanced 10 nm process technology, it achieves an
industry-leading combination of low power and high performance in both
human vision and computer vision applications. In addition, CV2S also
includes dedicated hardware for the stereo processing unit that can support
stereo IP cameras.
The CV2S’s CVflow architecture provides cutting-edge computer vision
processing at full 4K resolution, with up to 20 times the CNN performance of
Ambarella’s CV1
®
to enable image recognition over long distances and with
high accuracy. It includes efficient 4K encoding in both AVC and HEVC
video formats, delivering high-resolution video streaming with very low
bit rates to minimize cloud storage costs. The CV2S’s next-generation
image signal processor (ISP) provides outstanding imaging in low-light
conditions while high dynamic range (HDR) processing extracts maximum
image detail in high contrast scenes, further enhancing the computer vision
capabilities of the chip. It includes a suite of advanced security features to
implement advanced on-device physical security, including secure boot,
TrustZone
®
, and key storage. A complete set of tools is provided to help
customers easily port their own neural networks on the CV2S.
Image Signal
Processor (ISP)
Video Codec
H.264
H.265
MJPEG
Quad Core
Arm
®
Cortex
®
-A53
Ambarella Memory System
Security
DRAM Scrambling, TrustZone, Random
Number Generator, Secure Boot, OTP
Video Input
HiSPi
™
Video Output
DDR
Interface
Audio Input
(I2S)
SD / SDIO NAND / NOR
SPII2C GPIO
Timers
WDT
Ethernet
LCD
10 / 100 /
GigE
LPDDR4 /
LPDDR4x
SD Card /
eMMC
Sensor 1
Sensor 6
MIPI CSI-2
Analog
Parallel
MIPI CSI-2
SLVS
LVDS
Parallel
HDMI
MIPI DSI
64-BIT
Stereo and Computer Vision Processor
USB 2.0
Host / Device
Wifi / BLE
JTAG
UART
Audio Output
(I2S)
评论0