International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 05 Issue: 08 | Sep 2018 www.irjet.net p-ISSN: 2395-0072

Development of an Autonomous Drone for Surveillance Application

Dinesh M.A.1, Santhosh Kumar S.2, Sanath J. Shetty3, Akarsh K.N.4, Manoj Gowda K.M.5
1Assistant Professor, Dept. Of E&CE, Mit Mysore, Karnataka, India.
2,3,4,5Student, Dept. Of E&CE, Mit Mysore, Karnataka, India.
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Drones/unmanned aerial vehicles (UAV) are catastrophe. While a manned aerial vehicle puts the pilots at
defined as an aerial vehicle that does not need human risk during the crash, a UAV is able to fly anywhere even into
intervention. Drones are best known for their use in military the zones where it would be dangerous for the pilot.
operations including the use of weaponized drones for
targeted killing, especially in pre-emptive strikes and 2. PROPOSED METHODOLOGY
surveillance at a specific location. The UAVs can be deployed
for military applications to counter both domestic and While designing the autonomous drone, the auto-
international terrorism. This paper describes the development pilot was used with four brushless DC motors, GPS module,
of a UAV for surveillance in the region of deployment for telemetry, and the receiver. The whole system was powered
monitoring, based on human body detection along with via a power module with a Lithium Polymer battery of 18v.
soldiers uniform color and pattern recognition. This has been When the system is powered up and necessary configuration
achieved by using most stabilized flight controller and sending and calibrations are done, the system is armed for the flight.

the live visual feedback to the monitoring station and mail Flight controller interacts with other units via
alert message to the control station using SMTP protocol and standard communication interfaces such as pulse width
OpenCV library for object detection. modulation (PWM) module[1].

Key Words: UAV, OpenCV, Raspberry Pi, Pixhawk, IOT. For the autonomous mission, Pixhawk controller is
the most preferred one. It is supported with a number of
1. INTRODUCTION flight modes and most of the flight modes associated with a
GPS signal. In ‘Return to Launch’(RTL) mode, flight
Recent development and advancement in the engineering controller will note down takeoff coordinate points so that in
field such as propulsion system, airframes, batteries, and case of an emergency such as low battery or poor signal
microcontroller are seen as an important issue. The concept strength, it will automatically return to the position from
of an unmanned aerial vehicle is of dire importance to take where the flight was started. During the arming process, the
this development phase to an entirely new horizon. auto-pilot checks the fail-safe conditions like low battery
Development of an aerial vehicle that can operate voltage, GPS unable to lock, recalibration of the magnetic
autonomously and perform a designated task on its own compass or any other sensors. When all the fail-safe test is
without the considerable guidance of a pilot is in itself a task passed, the drone is ready for take-off.
to achieve. UAV can be managed remotely via radio, but
currently, a new control system has been developed which is On the other side, the waypoints are set using "Mission
capable of maintaining stable UAVs in the air along the planner" a software which has a great graphic user interface
proposed route. This intelligent system is placed on board (GUI) and makes it possible to change the upcoming
with UAV to replace human factor in flight operations and is waypoints altitude, change the mode of flight and can even
called as an automatic pilot or autopilot. Thus they control change the direction of flight. The drone in flight and the
the altitude, speed heading and other variables of the UAV computer with the mission planner software is connected by
autonomously by generating required control signals. telemetry module using radio frequency (MAV-link
protocol).
The UAV can be operated by radio frequency controller and
send live visual feedback which can be controlled by The Raspberry-Pi which is powered up using a buck
onboard computers or by remote control of a pilot on the converter is connected to a computer in the base station
ground. This UAV technology can be deployed for military using open source remote desktop (XRDP) protocol.
application to counter both domestic and international Raspberry Pi camera module offers better image
border issues. This paper describes the surveillance using representation with reduced noise contamination in the
human body detection along with color pattern recognition. image. UAV sends a live visual feedback to the ground station
The proposed system notifies any unauthorized entry along and also sends mail alert notification in case of detection of
with GPS coordinates which aid to locate enemies. an any unauthorized entry using the SMTP protocol.

In case of remote-controlled drone, there might be a chance

of damage in transmission section which would result in loss
of control over the flight which in turn may lead to

© 2018, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 331
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 05 Issue: 08 | Sep 2018 www.irjet.net p-ISSN: 2395-0072

Python is widely used high-level language and

extensively used with Raspberry Pi. Python syntax allows to
express concepts in fewer lines and it support wide user-
defined class.

 OpenCV

Open Source Computer Vision (OpenCV) is a cross-

platform library using which one can develop real-
time computer vision applications. It mainly focuses
on image processing, video framing, and analysis
including features like face detection and object
detection[3]. OpenCV has a wide objective that
mainly used to convert image pixel into a grayscale
image and processes the image depending on the
intensity level of each pixel.

Fig -1: Drone Design

3. IMPLEMENTATION

Mission planner software developed by Michael Oborne

which was specially designed for autonomous flight mission.
The software is used to achieve communication between the
base station and UAV. It is also used to program the
commands such as returning conditions, GPS coordinates,
altitude points etc., into a drone. The software allows the
flight controller to fly using various flight modes such as
acrobatic mode, altitude hold, loiter mode and stabilize
mode. There are a number of calibrations to be done on the
flight controller in synchronization with other hardware
components using mission planner[2]. Few of the mandatory
hardware calibrations are briefed below.

 Compass Calibration

Electronic compass in the Pixhawk module is

affected by the magnetic field generated by the
motors and also other nearby magnetic fields. By Chart -1: System Flowchart
calibrating the compass, one can compensate the
Step1: Initialize and import the OpenCV, SMTP library which
effect of the surrounding magnetic fields and in
is extensively used in image processing into the python IDE
turn, the drone can fly in the determined path.
platform.
 ESC Calibration
Step2: Capture the data frame, in which OpenCV default
Electronic speed controller (ESC) is responsible for color space model is BGR format well known as the RGB
rotating the motors at the speed required by the model. Camera module captures continuously the frames
autopilot. ESCs need to be calibrated so that they and converts it to a grayscale image[4], so that template
know the minimum and maximum pulse position matching is easier in grayscale intensities than the RGB color
modulation (PPM) values that the flight controller model, which also enhances processing delay.
will send.
Step3: Using OpenCV objective cv2.blur( ), we can reduce
sharpening edges by convolving with a filter that assigns an

© 2018, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 332
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 05 Issue: 08 | Sep 2018 www.irjet.net p-ISSN: 2395-0072

average pixel intensity value of neighboring 3x3 filter size. 3. CONCLUSIONS

Meanwhile match the grayscaled data with cascade classifier,
for upper body, full body, frontal face detection. A UAV usually has a rapid launch as compared to manned
aircraft while it has maneuverability as compared to ground
Step4: If the UAV is at a higher altitude then it should vehicles. This system can communicate using the wireless
suspect using specified color, now we assign the lower and network to receive control instructions from the base station
upper-intensity range to find out the threshold value to track as well send the images taken from the UAV to the base
suitable color. station. New methods are being developed for data collection
and advanced image processing of remotely sensed data. The
Step5: Find the contour (which is a mapping of similar UAV can be deployed in building and ships surveillance with
intensity pixel value) of tracked color and adjust the area if it the use of a proportional integral derivative (PID) closed
exceeds the contour range. And encircle that tracked color to loop gain feedback controller.
locate in data frames.
Using autonomous flight mission planning i.e., by setting
Step6: Check the condition, return true if data frame matches up waypoints with accurate geographical co-ordinates it can
with cascade classifier or if the specified color is identified. also serve as medicine carrying drone with first aid
treatment facilities which can be used in case of a medical
Step7: If the condition is true, mail alert message is sent by emergency.
UAV to notify the ground control station using SMTP
protocol[5]. REFERENCES

Step8: Clear and refresh the frame to capture the next frame [1] Mahmood ul Hassan, Muhammad Ahmer Ikram, and
for further processing until it is terminated. Dr.Syed Irtiza, “Development of an Autonomous Flight
Controller for Surveillance UAV,” IEEE 2016,
4. RESULT International Conference Emerging Technology (ICET),
DOI: 10.1109/ICET.2016.7813249.
An autonomous drone using effective utilization of GPS
mapping with flight controller has been developed. [2] Pawel Smyczynski, LukaszStarzec, and Grzegorz,
“Autonomous Drones Control System for Object
Tracking,” IEEE, 2017, 22nd International Conference on
Methods and Models in Automation and Robotics,
DOI:10.1109/MMAR.2017.8046919.

[3] Hani Hunud A. kadouf, Yasir Mohd Mustafah, “Color-

Based Object Detection and Tracking for Autonomous
Quadrotor UAV,” IOP Conference Series on 5th
International Conference on Mechatronics (ICOM13),
DOI: 10.1088/1757899X/53/1/012086.

[4] Tussanai Parthornratt, Natchaphon Burapanonte,

“People Identification and Counting System Using
Raspberry Pi,” IEEE, 2016, International Conference on
Fig -2: Autonomous Drone Flying
Electronics, Information, and Communication, DOI:
Live frame visual feedback is sent from drone to the 10.1109/ELINFOCOM.2016.7563020.
control station, which is used to detect, suspect and notify
[5] Ahmed AbdulQader Al-Bakeri, Abdullah Ahmed
any unauthorized human entries using body and color
Basuhail, “Notification System Based on Face Detection
detection algorithms.
and Recognition: A Novel Approach,” International
Journal of Computer Science and Information Security
(IJCSIS), vol. 14, pp187-191, April 2016.

[6] Akash Sehrawat, Tanupriya Choudhury, “Surveillance

Drone for Disaster Management and Military Security,”
IEEE, 2017, International Conference on Computing,
Communication and Automation (ICCCA), DOI:
10.1109/CCAA.2017.8229846.

Fig -3: Body and Color Detection

© 2018, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 333