Arducam Camarray Solutions

Expand a single CSI connector into several to connect and sync more MIPI cameras to your embedded systems like Raspberry Pi, NVIDIA Jetson Nano/Xavier NX, and more.

The 2nd Gen.Arduchip Stereo Camera Solution available in 3 styles and customization

Arducam’s breakthrough in simultaneously connecting more than one camera on regular Pi boards started in 2019 – when the Arducam stereo camera HAT is released to connect two cameras and synchronize them at clock level by I2C broadcasting and oscillator sharing.

The Camarray HAT is the second generation of Arducam multi-camera HAT solution. It adds support for RPi 12MP IMX477 HQ camera, and has stereo-camera and quad-camera HAT variations. It now supports global shutter cameras for machine vision.

3 arducam camarray HAT second generation update

Stereo Camera Module Style

Featuring:

Stereo Camera HAT Style

Featuring:

Quad Camera HAT Style

Featuring:

Support Raspberry Pi, Jetson Nano/Xavier NX, and more

As long as the MIPI camera module already works on those popular embedded systems. Arducam Camarray makes multiple of them possible.

raspberry pi logo 01

MMAL

Support native raspistill and raspivid commands on V1, V2, and HQ camera on the Raspberry Pi. Namely OV5647, IMX219, and IMX477.

nvidia logo 01

NVARGUS

Support native nvarguscamerasrc commands with IMX219 and IMX477 camera Modules on Jetson Nano and Jetson Xavier NX.

linux logo 01

V4L2

Support Arducam V4L2 camera drivers that make more image sensors – such as OV9281 and OV2311 – possible for the Raspberry Pi and Jetson platform.

B0267

More Connectors, More Possibilities

Arducam Camarray HAT solution allows you to connect more than one camera to a single MIPI CSI-2 camera interface. For example, it’s now possible to connect and sync up to 4 camera modules on a standard Raspberry Pi.

If you have more MIPI interface available on your processing board, the Arducam Camarray solutions are able to bring you even more possibilities to interface multiple cameras. Let’s say your Jetson carrier board comes with 2 MIPI CSI connector – we can quadruplex each of them and help interface up to 8 cameras on a standard Jetson carrier board.

The Arducam Community got your back

Here at Arducam, we are trying our best to provide the most complete support with the help of our community. You can read about our documentations about the Camarray solution on our website, and exchange your ideas on our forum.

If you’ve encountered an issues that are really urgent or difficult to handle, you can always contact Arducam specialist and reach out to us at [email protected]

Documentations

Documentations like quick start guides and datasheets are available.

Forum

Join the discussion in the Arducam forum to receive voices from us and other ​users.

Technical specialist

Our camera specialist can help setup a remote debug to narrow them down.

Give and Take

While the computing power of the embedded system remains the same, we have some compromises to make when getting more cameras running on more interfaces.

bg 01

Give – What are the limits we have to live with

Limited resolution and framerates with multiple cameras connected via a single CSI connector.

Those embedded systems come with limited camera interfaces for a reason – the system itself has a limited computing power to process the camera work. In other words, the total bandwidth allocated to the camera feed is limited. Therefore, more cameras require fewer frame rates and lower resolutions to work.

bg 02

Take – What are the limits we can break

You get to view and utilize the video feed from more cameras.

The Raspberry Pi 4, for example, only offers a single CSI camera connector. Admittedly, the Raspberry Pi may not be powerful enough to process multiple cameras simultaneously, but that should not exclude us from exploring its possibilities. With Arducam Camarray solutions, we help you expand the physical connectivities so you can build more and better projects.

Important Note on the Arducam Camarray HAT:

The Camarray HAT is not sold alone for the following reasons:

  1. Most of the cameras on the market are not designed for synched stereo uses. We’ve modified our camera modules specifically for the Camarry HAT. Your own V1, V2 or HQ camera modules won’t directly work with it.
  2. The Camarray HAT requires different firmware for different Image sensors that are not natively working with your embedded systems. Your own OV9281 and OV7251 camera modules won’t directly work with it, and they are not easily interchangeable.

Customization is required if you have extra requirements as the following:

  1. You need the camera module to run on other Video Modes and Frame Rates.
  2. You need to trigger the camera with an External Hardware trigger.

IMPORTANT NOTE ON THE IMX477 MIPI LANES:

Our IMX477 camera module support 4-lane in hardware, but the 4-lane camera driver requires extra customization. Here is why:

  1.  Standard Raspberry Pi and Jetson Nano/Xavier NX can’t physically support a 4-lane connection (both of them only have 2-lane CSI connectors). Only the Raspberry Pi Compute Module and some 3rd party Jetson Carrier Boards come with 4-lane hardware support. Therefore, it’s not a general need.
  2. Extra works, specifically camera drivers, are required for a 4-lane connection other than the 4-lane physical connector. And the drivers are limited by the performances of the platform itself. For example, 4-lane can’t unlock 4k video on Raspberry Pi because the VideoCore does not currently support it. Jetson Nano is also to some extent limited because of the encoding capabilities. You need to make sure you really need it.

More in the Blogs

Arducam Blog Thumbnail camarray release

Introducing Camarray: Arducam’s New Solution to Multiple Cameras on Embedded Systems and Edge Devices for 2021

Preface As one of the leading companies that pioneers in multiple cameras for popular embedded systems, Arducam now releases the new multi-camera solutions – named Camarray – which allows you to connect and sync up to 4 MIPI camera modules on the Raspberry Pi, Jetson Nano, and Jetson Xavier NX […]

12 comments