Skip to Content
Find More Like This
Return to Search

Image Processing Occupancy Sensor

National Renewable Energy Laboratory

Contact NREL About This Technology


PDF Document PublicationIPOS Fact Sheet (485 KB)

Occupancy detection--stationary and active
Occupancy detection--stationary and active

Multiple independent lighting and ventilation controls
Multiple independent lighting and ventilation controls

Technology Marketing Summary

Lighting controls offer significant potential for reducing that energy use, and new technologies that have emerged in recent years have enabled a wide range of innovative strategies, from room-level awareness of occupancy and daylight sensing to building-wide coordination of a fully networked system. About 40% of the average facility’s electrical load is dedicated to lighting, creating a sizeable market for energy efficient lighting hardware and software tools. The global market for intelligent lighting controls will experience steady and robust growth over the rest of this decade, rising from $1.5 billion in 2012 to more than $4.3 billion in 2020.

There are various lighting controls employed in the industry today. These include; occupancy, daylighting, personal tuning, institutional tuning, and integrating multiple control types. Analyses of individual lighting controls save on average between 25%-33%, while combined controls strategies can capture up to 40% savings on average. Occupancy sensing has remained an element of advanced building automation systems (BAS) for decades. Infrared or ultrasonic sensors are typically used to detect motion in a zone, providing feedback regarding occupancy or vacancy to the BAS or energy management system (EMS). Problems arise from these traditional methods of inferring building occupancy using infrared or ultrasonic sensing elements perform poorly under a variety of circumstances.  Even with careful tuning, false positive or negative signals cause a high degree of occupant annoyance, limiting their use for set point generation or daylighting.


NREL scientists have demonstrated unique image processing occupancy sensor (IPOS). The sensor is based on commercially available embedded hardware widely used by the smart phones industry, leverages mature open-source computer vision software libraries, and uses modern software engineering techniques and firmware development technologies.

IPOS overcomes several issues related to traditional passive infrared and ultrasonic-based motion sensors commonly used as proxies for direct occupancy detection, and provides a rich set of feedback signals with more sophisticated applications for use in occupant optimized lighting, daylighting, temperature, and ventilation control.  This technology increase accuracy of detection 20% better than current technologies and offers increased efficiency with the use of adaptive time delays. The small form factor, low-cost powerful processing hardware, and Linux base software make the IPOS highly cost and use effective.

The IPOS used a synergistic approach with face and people detection through computer vision algorithms. The IPOS classifies additional information including the number of occupants and their positions on polar coordinates, activity level (sedentary/active) and communicates this information with building automation systems via standard protocols (BACnet, others). The IPOS is developed and ready for testing implementation. 

  • Energy efficiency
  • Cost reduction
  • Increased accuracy
  • Small form factor
  • Multipurpose
Applications and Industries
  • Commercial and residential buildings
  • Lighting controls
  • Occupancy sensors
  • Security sensors
Patents and Patent Applications
ID Number
Title and Abstract
Primary Lab
Patent 9,036,866
Image-based occupancy sensor
An image-based occupancy sensor includes a motion detection module that receives and processes an image signal to generate a motion detection signal, a people detection module that receives the image signal and processes the image signal to generate a people detection signal, a face detection module that receives the image signal and processes the image signal to generate a face detection signal, and a sensor integration module that receives the motion detection signal from the motion detection module, receives the people detection signal from the people detection module, receives the face detection signal from the face detection module, and generates an occupancy signal using the motion detection signal, the people detection signal, and the face detection signal, with the occupancy signal indicating vacancy or occupancy, with an occupancy indication specifying that one or more people are detected within the monitored volume.
National Renewable Energy Laboratory 05/19/2015
Patent 9,456,183
Image processing occupancy sensor
A system and method of detecting occupants in a building automation system environment using image based occupancy detection and position determinations. In one example, the system includes an image processing occupancy sensor that detects the number and position of occupants within a space that has controllable building elements such as lighting and ventilation diffusers. Based on the position and location of the occupants, the system can finely control the elements to optimize conditions for the occupants, optimize energy usage, among other advantages.
Technology Status
Technology IDDevelopment StageAvailabilityPublishedLast Updated
NREL ROI 10-09 and ROI 12-66PrototypeAvailable - This technology is under option.11/30/201211/08/2012

Contact NREL About This Technology

To: Erin Beaumont<>