People counting algorithm ensures efficient space utilisation

People counting algorithm ensures efficient space utilisation
Technology News |
Algorithm provides system edge node analytics that unlock insights to improve space utilisation, people safety, and energy efficiency of spaces.
By Jean-Pierre Joosting

Share:

Analog Devices, Inc., (ADI) has introduced the EagleEye™ ADSW4000 PeopleCount algorithm for people detection and count in indoor spaces such as meeting rooms or office cubicles. The ADSW4000 is the first in a series of application-level software algorithms forming part of the ADI EagleEye platform that includes a hardware subsystem based on ADI’s Blackfin® embedded digital signal processor (DSP) and a set of application-level software building blocks enabling users to quickly develop their own people counting system. The ADSW4000 algorithm provides system edge node analytics that unlock insights to improve space utilisation, people safety through distance monitoring, and energy efficiency of spaces within intelligent buildings.

The ADI EagleEye platform enables a complete, end-to-end hardware and software system to help solve occupancy challenges. Combining the proprietary ADI ADSW4000 algorithm and advanced Blackfin DSP enables efficiency of indoor space while maintaining security and privacy as images are processed but not transmitted from the edge node. From meeting rooms to cafeterias, and lobbies to open desk people counting, ADI EagleEye can solve workspace needs and maximise worker wellbeing and productivity.

Key features include:

  • People counting algorithm for indoor areas, such as meeting rooms or office cubicles.
  • Occupancy state.
  • Location of person (x and y coordinates).
  • Zoned performance (multiple areas for larger coverage).
  • Edge processing.
  • Metadata output over UART.
  • No captured images are transmitted.
  • Coverage area for optimal accuracy of 90%: 3 m radius.
  • Maximum coverage area (accuracy of 80%): 5 m radius.

ADSW4000 data sheet

Linked Articles
eeNews Embedded
10s