Raspberry Pi has recently announced its new AI camera, which is capable of processing images directly on the device without the need for a GPU or hardware accelerators. In collaboration with Sony Semiconductor Solutions Corporation (SSS), Raspberry Pi has developed a new AI camera that integrates the IMX500 sensor, enabling image processing with AI models directly on the device’s processor. This eliminates the requirement for high-end GPUs or cloud computing, providing a more straightforward and cost-effective solution for users interested in deploying AI technology.
Priced at $70, the Raspberry Pi AI Camera targets developers in need of a powerful yet accessible AI tool. Compatible with single-board computers and Raspberry Pi microcomputers, this camera allows users to utilize familiar hardware to deploy AI-based image solutions without the need for specialized hardware devices. With a resolution of 12.3 megapixels, the camera supports video recording at 10 fps with a resolution of 4,056 x 3,040 or 40 fps at 2,028 x 1,520. Notably, it can run AI image processing tasks directly on the camera’s firmware through its integrated neural network, reducing the load on the host device like Raspberry Pi.
By processing data directly on-site, the camera optimizes performance, shortens processing time, and saves energy, eliminating the need for internet connectivity and large computational resources. Additionally, it is compatible with various hardware and software within the Raspberry Pi ecosystem, including popular libraries like lib-camera and Picamera2, making it more accessible to developers familiar with these tools. Overall, the Raspberry Pi AI Camera offers a low-cost solution for developers aiming to create AI systems without the complexity of specialized hardware devices.