Real-time object detection system for building energy conservation : an IP camera based system

Chandrasiri, AP; Hemba Geekiyanage, MD

Abstract

In the contemporary world, there is a rapid introduction of automated and intelligent
building systems. These technologies offer new and exciting opportunities to increase
the connectivity of devices in built environments, particularly for energy
conservation. Most of the developed building energy conservation systems are based
on sensors, thus, the application of those systems is limited to small spaces due to
maintainability issues. The reliability of these sensor-based systems is still argued as
sensors are not capable enough for multi-person tracking and real-time object
detection. Giving emphasis to these limitations, the current study introduces a realtime object detection, tracking and counting system for building energy conservation
particularly, for HVAC and lighting based on IP CCTV cameras. An experimental
research design was employed for the study. Initially, CCTV images from three
objects: human heads, lighted vehicles, and non-lighted vehicles were collected from
12 offices. Subsequently, these objects were trained using the machine learning and
the real-time object detection was performed using a Single Shot Detector model.
The proposed system was developed using the Python programming language. The
developed system comprised of three basic features namely, object detection, object
tracking and counting, and HVAC and lighting control. This system enables real-time
object classification for human heads, lighted vehicles, and non-lighted vehicles,
therefore, reduces excessive energy consumed by air conditioning and lighting
depending on the nature and movements of the objects. With the use of this system,
facility managers can make built environments much comfortable for occupants while
deducting excessive energy consumption and human effort taken to manage comfort
levels of buildings.

Citation

Chandrasiri, A., & Hemba Geekiyanage, M. Real-time object detection system for building energy conservation : an IP camera based system. Presented at 34th Annual Association of Researchers in Construction Management Conference, Belfast, UK

• Presentation Conference Type: Other
• Conference Name: 34th Annual Association of Researchers in Construction Management Conference
• Conference Location: Belfast, UK
• Publication Date: Sep 5, 2018
• Deposit Date: Oct 12, 2022
• Publisher URL: http://www.scopus.com/inward/record.url?eid=2-s2.0-85055626405&partnerID=MN8TOARS
• Additional Information: Event Type : Conference