Publication: Lighting control system based on human detection
Loading...
Date
2024
Authors
Tan, Yee Wei
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Nowadays, optimizing energy consumption is important as there is a limitation of energy production in industrial sector especially during increasing of human
population in the world. This project is dedicated to create an energy efficient automated smart lighting system for manufacturing floors to adjust lighting based on real-time human presence by integrating with two CCTVs and relay to turn on or turn off four lamps in the laboratory. The project is progressed in three stages. First stage is to get the RTSP address into the Deepstream pipeline. Second stage is to develop the Deepstream pipeline which consists of human detection model to detect the presence of human from real time video source via image segmentation method. Third stage is to send the output signal from the pipeline to the relay switch to control the lighting. The work applies the machine learning concept to integrate the real time video source from IP cameras and relay switches to control the lighting with the Deepstream pipeline. PeopleNet from TAO pretrained model is used to detect the presence of human from the video source via the image segmentation technique. The result obtained showed that the model can perform well at the accuracy of 100% in fully bright environment which is 320 lux. Finally, the main objective for this project is to apply the system at the manufacturing floor if the project is developed successfully.