Introducing a potent new AI camera for developers, Sony and Raspberry Pi

Together, Raspberry Pi Ltd. and Sony Semiconductor Solutions (SSS) have unveiled a brand-new device for fans of artificial intelligence. By analyzing visual data at the edge, the Raspberry Pi AI Camera, which works with all Raspberry Pi single-board computers, simplifies the process of developing AI solutions. The cost of the AI Camera is now $70.

This partnership stems from SSS’s April 2023 minority investment in Raspberry Pi Ltd. Since then, the businesses have collaborated to develop an edge AI platform specifically for the developer community of Raspberry Pi. The AI Camera’s primary function is the intelligent vision sensor (IMX500) from SSS. With the help of this cutting-edge sensor, AI processing can now be done right on the chip, doing away with the requirement for extra hardware like GPUs or accelerators, which is often needed to handle enormous amounts of visual input.

All of Raspberry Pi’s single-board computers, including the recently released Raspberry Pi 5, are compatible with the AI Camera. By using well-known tools like the libcamera and Picamera2 libraries, developers may streamline and expedite the integration process.

“As the first fruits of our strategic partnership, I’m very excited to share SSS edge AI sensing technology with the world’s largest development community,” said Eita Yanagisawa, general manager of SSS’s System Solutions Division.

The CEO of Raspberry Pi Ltd., Eben Upton, continued, “Together with our longstanding image sensor partner Sony Semiconductor Solutions, we have developed the Raspberry Pi AI Camera, incorporating Sony’s image sensor expertise. We look forward to seeing what our community members are able to achieve using the power of the Raspberry Pi AI Camera.”

The 12.3-megapixel IMX500 sensor that powers the Raspberry Pi AI Camera can take high-resolution pictures at 10 or 40 frames per second, depending on the setting. It is pre-loaded with the MobileNetSSD model and has an RP2040 microcontroller for managing neural network firmware, allowing it to be used right out of the box.

 

Komal Patil: