4K USB AI camera is powered by Rockchip RV1126 SoC with 2.0 TOPS NPU

ThinkCore TC-RV1126 is a USB AI camera powered by Rockchip RV1126 quad-core Cortex-A7 processor with a 2.0 TOPS NPU and equipped with a Sony Starvis IMX415 CMOS image sensor with 4K resolution (3840 x 2160).

The Linux-powered camera module features a USB Type-C port compatible with UVC & UAC protocols so that it can work without drivers on Windows, Android, Linux, Mac OS, and most other operating systems with a USB stack.

TC-RV1126 4K USB AI camera

TC-RV1126 camera specifications:

  • SoC – Rockchip RV1126 quad-core Arm Cortex-A7 @ 1.5 GHz + RISC-V MCU @ 400 MHz, 2D graphics engine, 2.0 TOPS NPU with support for INT8/ INT16, 14M ISP
  • System Memory – 2x DDR3 chip (128M x 16), so that should be 512MB RAM (TBC)
  • Storage – 1 Gbit (128MB) SPI NAND flash
  • Camera
    • Image sensor – Sony Starvis IMX415 1/2.8-inch 8MP CMOS sensor with night vision
    • Capture up to 3840×2160 @ 30 fps
    • Supported resolutions – 3840×2160, 1920×1080, 1600×896, 1280×720, 960×540, 640×480, 640×360, 320×240
    • Image output format – YUV, MJPEG, “Smart H.264”, “Smart H.265”
    • Fixed focus
    • Effective focal length (EFL) – 3.38mm
    • Aperture – F2.2
    • Field of view – Diagonal: 85.8°, horizontal:77.5°, vertical:48.8°
    • Optical structure – 6G + IR
    • Distortion – ≤17.8%
  • Microphone – Stereo microphone with up to 5m range
  • USB – 1x USB Type-C port for communication with host and power; support for UVC 2.0 for video and UAC 1.1 for audio
  • Power Supply – 5V/1A via USB Type-C port
  • Dimensions – TBD

4K USB AI camera

We told the “Tools, Driver and Firmware” can be found at the following link, but there’s nothing there except a link to Youyeetoo support forum with no reference to RV1126 at all, and an email address. So they may only be available to customers.

Applications for the camera include online teaching, live broadcast, video conference, video chat, surveillance, and more. The company goes on to list some of the supported programs like Tiktok, QQ, WeChat, Skype, and many more, but since it’s a standard USB camera it will just work like a webcam. That’s unless you’d like to make use of the AI features, in which case a specific program would have to be used, or at least an API to integrate the camera into your own application. ThinkCore does not provide public information, but if we look into the wiki for the Firefly CAM-C1126S2U camera module, it’s possible to use the RKNN toolkit, Rockchip’s rockface facial recognition program, or an SDK from ArcSoft.

ThinkCore TC-RV1126 4K USB AI camera is sold on Aliexpress for $69.82 including shipping. It’s unclear whether the plastic enclosure is included or not. People interested in purchasing in volume may want to check out the product page on ThinkCore’s website.

Share this:

Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress

ROCK 5 ITX RK3588 mini-ITX motherboard
Notify of
The comment form collects your name, email and content to allow us keep track of the comments placed on the website. Please read and accept our website Terms and Privacy Policy to post a comment.
1 year ago

I wonder if these kind of camera’s, see also pinecube, can be used with openscan, to have a all-in-one solution with the controller and camera in the same device.

Jon Smirl
1 year ago

Firefly has had a similar camera for over a year. But it is more expensive and not as good of sensor chip. Firefly has put all of the software up on github, Theirs is dual sensor to do facial depth mapping .

They claim working database of 10,000 faces with processing time under 30ms. Their application is building security.

Khadas VIM4 SBC