TSPS (Toolkit for Sensing People in Spaces) is an open-source, cross-platform computer vision library developed by Lab at Rockwell Group designed to easily track human presence, position, and movement within a physical environment. It is highly favored by digital artists, exhibition designers, and developers because it removes the need to write complex computer vision algorithms from scratch. Instead, TSPS functions as a standalone application that processes video feeds and instantly translates real-time human movement into clean data that other software can understand. How TSPS Works
TSPS acts as a bridge between the physical world and digital creative environments:
Input Sensing: You connect a visual sensor—such as a standard web camera, a Microsoft Kinect, a depth camera, or even a pre-recorded video file.
OpenCV Processing: Using OpenCV under the hood, the toolkit applies background subtraction and contour analysis to isolate people (“blobs”) from the environment.
Data Communication: Instead of rendering graphics itself, TSPS packages tracking statistics—such as an individual’s location coordinates (x, y), velocity, bounding box, and dimensions—and broadcasts them over common network protocols. Supported Communication Protocols
TSPS is exceptionally versatile because it transmits data seamlessly via several popular communication methods:
OSC (Open Sound Control): Ideal for sending data to interactive audio/visual software.
TUIO: A standard protocol for tangible multi-touch surfaces and gesture tracking.
Websockets: Perfect for streaming motion data directly to modern web browsers.
TCP/IP: Useful for robust, standard network-based programming. Compatible Client Ecosystem
Because it uses standard protocols, you can pick your favorite programming environment to build the actual user experience. TSPS comes with standard client examples for Processing, JavaScript, C++ (OpenFrameworks), Python, TouchDesigner, Unity, and Max/MSP. Primary Use Cases
Interactive Installations: Making a digital wall or projection floor react when a person walks past or steps on a specific area.
Exhibition & Retail Analytics: Anonymously studying how people navigate through museum exhibits or retail spaces to optimize spatial design.
Rapid Prototyping: Quickly building proof-of-concept interactive spaces without spending weeks configuring deep-learning models or camera calibrations.
Are you planning to build an interactive installation, or are you looking to integrate TSPS with a specific programming environment like Processing or TouchDesigner? Let me know your goals and I can guide you through the setup process!
Leave a Reply