Toy Vision-Guided 3D Robotic Arm in JavaScript

Modern JavaScript has the power to access the capabilities of the underlying architecture in a simple and portable manner and without installing third-party libraries. A one-page web application was built that connects to the camera, processes the video stream, extracts images, and detects motion. The detected motion is then used to control a 3D robotic arm rendered using webGL within the same page. Although the scope of this paper is purely didactic, the framework can easily be extended and used in a classroom setting for experimentation with different vision algorithms.