ON Semiconductor has announced the launch of the RSL10 Smart Shot Camera, a new device that aims to bring computer vision capabilities to ultra-low-power edge camera devices.
"Hyper automation is the next evolutionary step in the IoT," claims Wiren Perera, head of Internet of Things at ON Semiconductor. "The RSL10 Smart Shot Camera embraces that concept, by bringing the power of vision based AI closer to the edge to enable capabilities such as automatic object recognition and scene changes. We have been consistently delivering on a variety of device-to-cloud wireless sensor solutions. Enabling the seamless addition of vision takes this to another tier."
The RSL10 combines two key previous releases into one platform: The RSL10 system-in-package (SIP), which includes an Arm Cortex-M3 core running at up to 48MHz, 76kB of program and 88kB of data memory, 384kB of flash memory, and a 32-bit Dual Harvard DSP core along with Bluetooth 5 connectivity; and the ultra-low-power ARX3A0IAS miniature CMOS imaging sensor, which boasts a 360 frames per second (FPS) framerate at 560x560 resolution.
The RSL10 platform is designed to allow developers to create an endpoint with automatic image upload upon local detection of a change — whether that's triggered by time, light levels, temperature, or something else. In its low-power mode, the system can also look for motion and trigger an image capture and upload for analysis automatically. In either case, the upload takes place over Bluetooth Low Energy — and configuration is handled by a companion smartphone application.
ON Semi says the part is ideal for bringing cloud-driven artificial intelligence to everything from robotics to wearables, using the suggestion of a camera-equipped hard hat or a shopping cart which monitors items for automated checkout. The company has even suggested embedding the camera into cupboards to automatically fill out shopping lists as items are used.
More information on the RSL10 is available on the ON Semi website.