Publication: A Non-Intrusive Approach for Indoor Occupancy Detection in Smart Environments
| dc.contributor.author | Azreena binti Zaharil Azlan | en_US |
| dc.date.accessioned | 2023-05-03T16:27:28Z | |
| dc.date.available | 2023-05-03T16:27:28Z | |
| dc.date.issued | 2020-02 | |
| dc.description | FYP 2 SEM 2 2019/2020 | en_US |
| dc.description.abstract | Today, the automation of everyday tasks is targeted with smart environment. Smart environment comprises of smart objects that can connect with each other, share information, and coordinate actions in order to take smart and cognitive decisions according to the environment context. This has been made possible with technological advances of external sensors, coupled with computational processing. Smart objects are able to acquire occupancy information. Occupant information is a major input for context-driven control approach in smart environments. However, dealing with dynamic and frequent context changes is a meticulous task. Deep learning support is acquired for real-time data acquisition and processing platforms. This project provides solution for smart environment in classrooms with the objectives to design and model the classroom occupancy detection system, to analyse and fine tuning the real-time occupancy detection, and to investigate the developed real-time system in classroom. This project implements TensorFlow Object Detection API in Python programming for the execution of Convolutional Neural Network (CNN) and OpenCV module to provide real-time domain for the API. The codebase from TensorFlow Object Detection API is revised to have the feature that suits real-time occupancy detection. The approach of the project is obtaining real-time object detection. Next, changing of initialized ID class of object detection to acquire real-time human detection and finally achieving real-time human detection and counter. The output of this project highlights human detection and human counter in classrooms and can be implemented with both image and video input in real-time domain. | en_US |
| dc.identifier.uri | https://irepository.uniten.edu.my/handle/123456789/21285 | |
| dc.language.iso | en | en_US |
| dc.subject | Indoor occupancy detection | en_US |
| dc.title | A Non-Intrusive Approach for Indoor Occupancy Detection in Smart Environments | en_US |
| dspace.entity.type | Publication |