A Smart Security Surveillance System
Person detection and tracking is one of the most important research field that has gained a lot of attention in recent years. There are lot of surveillance cctv cameras installed around us thus, it is necessary to develop a computer vision based technology that automatically processes the real time video frames in order to track the person. Here, we aim to build an intelligent system which has its own secure database system, a cctv network and a network of connected nodes(systems). Each node will be connected to the centralized system. Node will be used by authorized personnel with their login credentials for their authorized operations in the system(searching,tagging,etc)
It will be installed on a single centralised machine where the main algorithm will be running and the database is accessed. This will be connected to the nodes from which the authorized personnel will get access to the system.
The algorithm will detect humans in the cctv footage frames and will search the database for identification of people in the live video footage. If there is one person of interest in the live feed, there will be an option to select that individual and only track activities of that person. After tracking, the algorithm will upload the logs of tracking to the database, in the tracked person/people’s records.
3. Software system for each node: To be installed in individual systems or ‘nodes’. Here, authorized personnel will perform operations. As per login credentials, access rights will be defined for the user.
CCTV surveillance system provides the real-time recording as well as online access by the staff to monitor more intelligently that would eventually reduce the risk of crime. CCTV surveillance system provides the real-time recording as well as online access by the staff to monitor more intelligently that would eventually reduce the risk of crime.
The major part of this project is done on Tensorflow. The aquisition of video is done using OpenCV and the GUI is produced using PyQt4. To run this project, the system is only required to have a Python 3.6.x interpreter installed.
For realtime performance, Tensorflow GPU is also required. Further details on Tensorflow.org
Detailed system requirements are produced in requirements.txt
After cloning the repository, at root directory :
pip install -r requirements.txt
The model files and some ommited google api keys are not present in this repository to limit usage of the package. For access, kindly contact one of the contributors.
For Linux Systems :
python3 gui.py (for GUI version)
python3 yolo.py (for non-GUI version)
For Windows Systems :
python3 gui.py (for GUI version)
python3 yolo.py (for non-GUI version)
- Tensorflow - The DL framework used
- Python - Language used
- PyQt - Used for producing GUI
- Opencv - Used for video aquisition
This project is licensed under the GNU GPLv3 License - see the LICENSE.md file for details