Open source hardware provider DFRobot has launched an AI-powered vision sensor, known as HuskyLens, that’s capable of detecting faces, objects, and colors with a push of a button, and continues to learn, and improve its abilities with each click of the camera.
“HuskyLens is designed to be smart. Built-in algorithms allow HuskyLens to learn new objects, faces, and colors through a single click. Moreover, by long pressing, HuskyLens can continually learn them even from different perspectives and in various ranges. The more it learns, the more accurate it can be.”
Those AI algorithms outfit the camera with facial recognition, object recognition/tracking, color detection, line following, and tag detection. Users can also select different algorithms by pressing some buttons, so even people with limited understanding of AI can easily use the camera. You can even combine different algorithms, enabling a small robot, for example, to follow a set line and determine which direction to follow.
The HuskyLens is based on the Kendryte K210 SoC with a dual-core RISC-V 64-bit processor designed with AI applications in mind, including machine learning, machine hearing, and convolutional neural networks. The camera is compatible with several popular boards as well — Raspberry Pi, Arduino, LattePanda, and micro:bit, to name just a few — allowing users to integrate the camera into almost any project.
DFRobot is currently crowdfunding the HuskyLens on Kickstarter, with pledges starting at around $20 for the standard early bird version, which gets you a 2-megapixel camera, and approximately $50 for an enhanced version with a 5-megapixel camera. At the $168 price point, DFRobot throws in a Devastator Tank Mobile Robot Kit with a Romeo BLE Arduino control board.