Aerostats and camouflaged object detection system made to propose a practical solution to identify and liquidate aerostats and camouflaged objects using a drone and a built-in laser module.
-
Aerostats object detection algorithm
The aerostats object detection algorithm is a Real-Time object detection algorithm, based on YoloV5 Pytorch implementation. The algorithm is capable of analyzing and inference from a remote drone camera, track the identified aerostat, and point a laser beam towards the target for its liquidation in real time, using a laser module that is installed on the drone, and is controlled by Servos and Arduino.
-
Camouflaged object detection algorithm
The camouflaged object detection algorithm is based on 13 Haralick Features, which Robert Haralick suggested in his article from 1973, and on a SVM (Support Vector Machine) model.
The algorithm is capable of analyzing and inference from images and videos. The objects that are predicted by the algorithm are visualized in a stream by mask object that is drawn inside the frame, and over the detected object.
Inside the project's root folder, 'Balloon-Camouflage-Detection', there is the main GUI component, called 'Demo_Dashboard'. Edit 'Demo_Dashboard' as follows:
-
In line 56, change first argument of 'subprocess.call' to the following command, after changing the red paths to the correct paths on your local machine:
[PYTHON_PATH]/python.exe [PATH_TO_FOLDER]/Drone-Balloon-Camouflage-RealTime-Detection/yolov51/detect.py --source 0 --weights [PATH_TO_FOLDER]/Drone-Balloon-Camouflage-RealTime-Detection/yolov51/best.pt --conf-thres 0.7 --half --dnn -
In 'detect.py' inside 'yolov51' subfolder, change line 51 and define the correct COM number for the Arduino USB interface:
ArduinoSerial=serial.Serial('COM6',9600,timeout=0.1) # Define COM number for arduino interface -
Install dependencies:
Make sure 'PIP' package is installed. Open PowerShell from 'yolov51' folder and execute:
pip install -r requirements.txt -
Run 'Demo_Dashboard.py', and choose the algorithm you want to fire:
