Embedded AI Vision Modules

Ambarella-Based AI Vision Camera Modules for Seamless Integration

Our AI vision camera solutions help to identify and track objects, analyse scenes, and predict behaviour in various industries: from video surveillance and security to agriculture and automotive software. We develop the right building blocks to create automotive cameras that collect visual data and provide valuable insights.

Promwad’s system-on-modules (SoM) are based on Ambarella CV25, CV22, H22, and H32 processors. These chips are known for low-power architecture combined with industry-leading AVC/HEVC encoding, making them suitable for various automotive vision systems and other Edge AI applications.

Aftermarket Dashcams
Security
Smart Home
Smart Agriculture
Action / Wearable Cameras
Connected Drones
Aliaksei Safonau
Discover how our team can tackle your biggest tech challenges!

Aliaksei Safonau, Head of the Automotive Unit at Promwad

Automotive AI Vision Systems: From MVP to Production

Automotive AI technologies have a primary focus on improving safety, driver comfort, and user experience. Leveraging our expertise in Ambarella-based SoM development, we help you create cutting-edge automotive vision solutions by:

  • Selecting the right components and boosting your concept with our Ambarella-based SoM ecosystem.
  • Assisting with requirements elaboration, providing project budget estimation, and performing a feasibility study in the early stages.
  • Executing your project from design to mass production.
  • Providing integration with various automotive solutions, such as embedded GPS trackers, ADAS, and EV charging software.
  • Manufacturing your products at a verified and state-of-the-art production site (EMS).
See how we can help you deliver results on time and within budget!

Ambarella Сhipset Сomparison and Target Applications

 

CPU Spec

4 x ARM Cortex-A53 @1GHz

4 x ARM Cortex-A53 @1GHz

4 x ARM Cortex-A53 @1GHz

4 x ARM Cortex-A53 @1GHz

AI / CNN

Software-based on ARM core

Software-based on ARM core

CVFlow processor with CNN / deep learning support 

CVFlow processor with CNN / deep learning support 

OS

ThreadX, Linux, ThreadX + Linux

Video encode

H.264(AVC)

H.265(HEVC)

H.264(AVC)

H.265(HEVC)

H.264(AVC)

H.265(HEVC)

H.264(AVC)

H.265(HEVC)

Video spec

Ultra HD 4Kx2Kp30

Ultra HD 4Kx2Kp30

Ultra HD 4Kx2Kp60

Ultra HD 4Kx2Kp45

Multi-stream

1440p + 1080p/720p

1440p + 1080p/720p

1440p + 1440p

1440p + 1440p

Video Input

2x

3x

3x

3x

Memory Interface

LPDDR4x / LPDDR4 / DDR4

32-bit bus width

LPDDR4x / LPDDR4 / DDR4

16-bit bus width

LPDDR4x / LPDDR4 / DDR4

32-bit bus width

LPDDR4x / LPDDR4 / DDR4

32-bit bus width

Interfaces

10/100/1000 

Gigabit Ethernet MAC


2 logical VOUT channels


I2S audio


USB 2.0 Host/Device with a built-in PHY


NAND flash controller


2x SD controllers


3xUART

10/100/1000 

Gigabit Ethernet MAC


2 logical VOUT channels


I2S audio


USB 2.0 Host/Device with a built-in PHY


NAND flash controller


3x SD controllers


5xUART

10/100/1000 

Gigabit Ethernet MAC


2 logical VOUT channels


I2S audio


USB 2.0 Host/Device with a built-in PHY


NAND flash controller


3x SD controllers


5xUART

10/100/1000 

Gigabit Ethernet MAC


2 logical VOUT channels


I2S audio


USB 2.0 Host/Device with a built-in PHY


NAND flash controller


3x SD controllers


5xUART

Target market

Connected drones, action cameras, wearable cameras

Aftermarket dash cameras, high-performance action cameras, wearable cameras

Intelligent sports, drone, wearable, and panoramic cameras

Intelligent sports, drone, wearable, and panoramic cameras

 

Do you want to employ our modules with advanced image processing and computer vision on your device?

Drop us a line about your project! We will contact you today or the next business day. All submitted information will be kept confidential.