Advanced Human Detection and Visualization with ML and ToF Camera Integration for a Global Manufacturer
Overview
A global manufacturer of advanced automation devices sought a reliable technology partner to build a human detection and visualization solution for smart buildings and premises. The client aimed to leverage their Time-of-Flight (ToF) camera sensors to detect, analyze, and visualize human presence, enabling scalable deployment for their B2B customers in the building automation and robotics sectors.
Download Case Study
Challenges
Lack of a real-time, intelligent solution to detect and visualize human presence using ToF camera data
Difficulty in processing cloud data for accurate human detection and counting
Solution
- Developed Linux based Machine Learning application to interface with Time-of-Flight camera sensor
- Processed 3D point cloud data from the camera sensor data using Support Vector Machine algorithm and Histogram of Oriented Gradients feature descriptor to detect human presence and counting
- Integrated off-the-shelf 3D viewer in the application to display input frames supporting 10 human detection in one frame and user configurable detection zone & frame resolution
- Integrated the application in Robot Operating System (ROS2) to expand the opportunity scope in robotic market as well
Outcomes
- Integration of Robot OS has helped the client to reduce the entry barriers for the client’s product in to the robotics domain by minimizing the system integration efforts
