Here’s good news for individuals and companies wanting to incorporate AI into their product’s imaging capabilities. Sony Semiconductor Solutions and Raspberry Pi have teamed up to drop the Raspberry Pi AI Camera. This smart little device processes visual data right at the edge without needing a fancy GPU.
Thanks to Sony’s IMX500 intelligent vision sensor, it handles AI processing on its own, making your Raspberry Pi much smarter in no time. It’s compatible with every Raspberry Pi model, including the new Pi 5, and plays nicely with libcamera and Picamera2 libraries, so you can start tinkering immediately.
This 12.3-megapixel camera gives you crisp images at 10 fps or 40 fps at lower resolutions, with a 76-degree field of view and adjustable focus. It even comes pre-loaded with the MobileNetSSD model, so it’s ready to detect objects right out of the box. Plus, with built-in AI muscle, your Pi can process data like a pro without any extra hardware.
Here’s a rundown of the specifications:
- Sensor model: SSS’s approx. 12.3 effective megapixel IMX500 intelligent vision sensor with a powerful neural network accelerator
- Sensor modes: 4,056(H) x 3,040(V) at 10 fps / 2,028(H) x 1,520(V) at 40 fps
- Unit cell size: 1.55 µm x 1.55 µm
- 76 degree FoV with manual/mechanical adjustable focus
- Integrated RP2040 for neural network firmware management
- Works with all Raspberry Pi models using only Raspberry Pi standard camera connector cable
- Pre-loaded with MobileNetSSD model
- Fully integrated with libcamera
If you’re ready to dive into AI image processing with just your Pi and a camera, the Raspberry Pi AI Camera can be had as we speak for US$70 from approved resellers like Adafruit, PiShop US, and more.
Images: Sony Semiconductor Solutions.