Drone Detection Surveillance System
Unmanned Aerial Vehicles (UAVs) like drones are widely employed in photography, videography and more recently, for transportation purposes. On the other hand, drones can be utilized in malicious activities, e.g., monitoring and spying of strategically important defensive installations, exploiting so
2025-06-28 16:32:13 - Adil Khan
Drone Detection Surveillance System
Project Area of Specialization Artificial IntelligenceProject SummaryUnmanned Aerial Vehicles (UAVs) like drones are widely employed in photography, videography and more recently, for transportation purposes. On the other hand, drones can be utilized in malicious activities, e.g., monitoring and spying of strategically important defensive installations, exploiting someone’s privacy, smuggling and carrying out terrorist attacks. Our project Drone Detection and Surveillance System is focused on developing an autonomous system that can detect drones in the vicinity through audio signal processing and computer vision. The presence of drones will be detected by a Machine Learning based classification model, trained on a characteristic drone audio feature. The trained classifier will be able to classify between the environmental and drone sounds and therefore, will intimate the presence of drone in the area. After the detection, the drone will be localized through audio signal processing and then tracked via a camera mounted on a turret by using a SIFT feature-based tracking algorithm. The complete drone detection surveillance system will be comprised of microphones, cameras, and turret hardware. The effort will be made to build a computationally tractable and simple system that can detect the drone in real-time with the minimum response time.
Project ObjectivesProject objectives include:
- Drone detection through Machine Learning using audio features.
- Drone localization through audio signal processing.
- Drone tracking through video processing and using SIFT based tracking algorithm.
- Hardware implementation of the proposed system.
- Performance evaluation and testing of the proposed system in real-time
Method of analysis (Audio)
- Audio Dataset (Audio Based Drone Detection and Identification using Deep Learning)
- Extraction of features:
- Mel-frequency cepstral coef?cient, MFCC
- Linear Prediction Cepstral Coefficients, LPCC
- Zero Crossing Rate, ZRC
- Spectral Roll off
- Training of system using optimum classifier (SVM and KNN)
- Localization of drone by DOA
Method of analysis (Video)
- Initial point given manually
- Extraction of SURF features of the bounding box of drone
- Feature matching in next frame
- Creation of bounding box
- Drone tracking throughout the video
- Compilation of the frames with the tracked drone
The developed system in this project can be utilized for security purposes for providing surveillance of strategically critical installations and areas in Pakistan. At the moment, no such system is installed and generally, security personals are deputed over the buildings to monitor the outdoor activity. As humans do not have surveillance ability of 24/7 and the fatigue, sleep deprivation can result in a security breach, the proposed system can be installed at such locations to save the human resource and to increase the security.
Technical Details of Final DeliverableIn terms of hardware, the final product will be composed of two microphones, a high-resolution camera accessible through MATLAB, a turret, motors, and Arduino. The microphones will capture the sound based on the Distance of the Arrival of the sound. The features of the sound will be extracted and given to the classifier for detection. If a drone is detected, based on the Distance of Arrival, the azimuthal angle will be calculated and fed into the turret on which a camera is mounted. The motor drivers, operated by Arduino, will move the turret to include the flying drone in the frame since the camera has an only 180-degree view. Now the camera will capture the flight pattern of drone which will be fed into the algorithm for its feature extraction. The drone will thus be tracked, and the user will be notified of its presence in the vicinity.
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 76000 | |||
| Microphones | Equipment | 2 | 9000 | 18000 |
| Arduino Uno | Equipment | 1 | 1000 | 1000 |
| Camera | Equipment | 1 | 25000 | 25000 |
| Electric Motor | Equipment | 1 | 1000 | 1000 |
| Audio Card Interface | Equipment | 1 | 16000 | 16000 |
| Motor Driver | Equipment | 1 | 1000 | 1000 |
| Localization Turret | Equipment | 1 | 6000 | 6000 |
| Connecting wires | Miscellaneous | 1 | 1000 | 1000 |
| Integrated Circuits | Miscellaneous | 1 | 7000 | 7000 |