The Raspberry Pi AI Camera, a collaborative effort between Raspberry Pi and Sony, is now available for $70. This innovative camera module works seamlessly with all Raspberry Pi microcomputers and eliminates the need for additional accelerators or a GPU. The camera features Sony’s IMX500 image sensor, known for its onboard AI processing capabilities.
This allows Raspberry Pi users to develop edge AI solutions that can efficiently process visual data. With a 12.3-megapixel resolution, the camera can capture footage at either 10 frames per second in 4056 x 3040 resolution or 40 frames per second at 2028 x 1520 resolution. It also has a manually adjustable focus and a 76-degree field of view.
The camera’s compact size of 25 x 24 x 11.9mm makes it almost identical to previous Raspberry Pi camera models. The CEO of Raspberry Pi expressed excitement about the potential applications community members might create with this new technology. He emphasized the growing importance of AI-based image processing.
This development highlights Raspberry Pi’s commitment to enhancing its camera offerings. The company launched its initial 5-megapixel product in 2013. The collaboration with Sony marks a significant milestone in making advanced imaging technology more accessible to developers worldwide.
The Raspberry Pi AI Camera introduces AI capabilities to Raspberry Pi users at an affordable price. With embedded AI technology, this camera offers significant potential for robotics, object detection, and other AI-driven tasks while maintaining low power consumption.
Introducing affordable ai imaging technology
The camera incorporates the Sony IMX500 “Intelligent Vision Sensor,” a 12MP camera with its own Neural Processing Unit (NPU). The onboard Raspberry Pi RP2040 acts as an interface, using an attached 16MB flash device to cache AI models. This allows it to perform AI tasks independently of the Raspberry Pi’s CPU, conserving power and freeing up the CPU for other tasks.
The AI camera resembles the Raspberry Pi Camera Module V3 but has a larger plastic shroud around the sensor. The mounting holes are consistent with previous models, making it easy to install in existing setups. Upon installation, the camera can begin its tasks.
It supports TensorFlow and can run pre-designed models to identify objects and poses. During testing, the AI camera maintained 30 frames per second for both object detection and pose recognition at a resolution of 640 x 480 pixels. Using the Picamera2 library, developers can integrate AI capabilities into Python projects.
Pose detection, for example, can trigger motors, sound effects, or other actions. The Raspberry Pi AI Camera is particularly suitable for educational environments, light industrial AI projects, and robotics. Its ease of use and integrated AI capabilities make it an excellent tool for learners and hobbyists.
In summary, the Raspberry Pi AI Camera combines affordability with robust functionality. It provides an excellent entry point for AI projects on the Raspberry Pi platform. While not as powerful as the AI Kit, it offers more than enough performance for most users, especially in educational and prototyping contexts.
