International Journal of Engineering Trends and Technology (IJETT) Volume 35 Number 4- May 2016

Object Tracking Robot on Raspberry Pi using Opencv

Pradeep Kumar.G.H
Assistant Professor, Department of CSE, KSIT

Rupa Gurram SweathaSuresh.B. Sneha.B.R. Sushmitha.R.

Students, Department of CSE, KSIT

ABSTRACT idea behind robotic vision is: viewing an object from the
on object of interest and to
Autonomous vision based robots are intelligent robots act accordingly. It is about giving artificial sight to
which take visual data, process it and provide appropriate robots.The implemented robot can be used in various
chemical industries, military application, coin separation,
need any kind of human intervention since they are pre- pencil industry.
fed with instructions. A robot is designed on Raspberry
Pi using OpenCV, which is used for object detection 2. HARDWARE DESIGN
based on its colour, size and shape. Here only a single
object is being detected at a time. The tracking of the 2.1 List of Hardware
object is based on division of the image into virtual grids.
The movement of the robot is based on the position of A chassis is used as a base on which following hardware
the object in the grid. An arm is incorporated to pick the components are fit:
object once it is stationary. After picking the object, it
would be dropped into its respective coloured container Raspberry Pi (rev B) for GPU and CPU
which has a predefined position. computations.
Wi-Fi dongle to connect to Pi remotely.
Keywords- OpenCV, Raspberry Pi, Virtual Grids, RDP Motor driver which can control two motors
Batteries to provide power.
Desk, GPIO, Contours. Jumper wires to connect individual
components.
1. INTRODUCTION Web camera.
Servo motor to control the arm of the robot.

According to the Oxford English Dictionary, the word 2.2 Hardware and Software Description
roboticswas first used in print by Isaac Asimov, in his
science fiction short story "Liar!". Asimov was unaware 2.2.1 Raspberry Pi
that he was coining the term and he assumed robotics
already referred to the science and technology of robots. The Raspberry Pi is a credit card-sized single-board
Stories of artificial helpers, companions and attempts to computer. There are currently five Raspberry Pi models
create them have a long history, but fully autonomous in market i.e. the Model B+, the Model A+, the Model B,
machines only appeared in the 20th century. the Model A, and the Compute Module. All models use
the same SoC (System on Chip - combined CPU&GPU),
Artificial Intelligence (AI) powers the evolution of but hardware features differ.
robotics in the modern era. The core of robotics is AI and
computer science is a building block of AI. Highly
intelligent robots and robots with AI are fruits of
continuous research and years of development. The
present project was time and resources bound to build a
robot with high level AI; hence, a simple, yet an
 International Journal of Engineering Trends and Technology (IJETT) Volume 35 Number 4- May 2016

quality and configurability of the camera module is 2.2.10 Wi-Fi dongle

highly superior to a standard USB webcam.
The Wi-Fi dongle is attached to the USB port in
2.2.3Raspbian OS Raspberry Pi in order to connect to it wirelessly. Since
raspberry pi needed its own IP, it needs to be connected
Raspbian OS is available for Raspberry Pi, Raspbian to a Wi-Fi router or Hotspot [7]. For the same we need to
comes out on top as being the most user-friendly, best- make some changes in the field specified so as to make
looking, has the best range of default softwares and raspberry pi recognize the router every time it boots up.
optimized for the Raspberry Pi hardware. Raspbian is a
free operating system based onDebian (LINUX), which following lines to make the PI connect with your router.
is available for free from the Raspberry Pi website.
iface wlan0 inetdhcp
2.2.4 Python wpa-ssid "Your Network SSID"
wpa-psk "Your Password"
Python is a widely used general-purpose, high-level
programming language .Its syntax allows the 3. ARCHITECTURE
programmers to express concepts in fewer lines of code
when compared with other languages like C, C++or java.

2.2.5RPi.GPIO Python Library

The RPi.GPIO Python library allows you to easily

configure and read-
GPIO header within a Python script. This package is not
shipped along with Raspbian.

2.2.6OpenCV

It (Open Source Computer Vision) is a library of

programming functions mainly aimed at real-time
computer vision.
Fig 2: Architecture diagram
2.2.7Numpy
A Live video is captured from the webcam. The pre-
Numpy is a package that defines a multi-dimensional processed frames are then converted to black and white ,
array object and associated fast math functions that object is identified by finding the contours of the object
operate on it.Numpy is an extension to the Python and then, centroid of the object is calculated. Pre-
programming language, adding support for large, multi- processing of the frames include noise removal and
dimensional arrays and matrices, along with a large conversion to grey scale. The frames are divided into
library of high levelmathematical functions to operate virtual grids which help the robot to move to the object
on these arrays. based on its centroid. Once the robot reaches the object
2.2.8 RDP Desk and finds the object is stationary for few frames, it picks
the object. The colour of the picked object is analysed
Remote Desktop Protocol (RDP) is a proprietary and then the object is placed into its respective coloured
protocol developed by Microsoft, which provides a user container.
with a graphicalinterface to connect to another computer
over a network connection. The user employs RDP client 4. QUADRANT BASED ALGORITHM
software for this purpose, while the other computer must
run RDP server software. Step 1: Capture live video.
2.2.9 Servo Motor Step 2: Pre-process the frames. This step includes
converting of the frames to gray scale and also removing
A servomotor is a rotary actuator or linear actuator that
 International Journal of Engineering Trends and Technology (IJETT) Volume 35 Number 4- May 2016

Step 8: Robot places the object into its respective

coloured container.

5. MODES OF OPERATION

5.1 AUTOMATIC MODE

The complete work of tracking and placing the object

into its respective coloured container is done by robot.

Fig 3 : grid

These grids are virtual and help in identifying the

position of the object.

Step 5: Object is identified by finding out its contour[6]

as shown in fig 4 which is found by the following.

contours1,_=cv2.findContours(im_bw,cv2.RETR_TR
EE,cv2.CHAIN_APPROX_NONE)

Fig 4: contour

Using contours we find the centroid of theobject, which

is calculated by the formula given below:

M=cv2.moments(im_bw)
centroid_x = int(M['m10']/M['m00'])
centroid_y = int(M['m01']/M['m00'])

By the use of the centroid we find the actual position of

the object(in which quadrant it lies).

Step 6: The main idea is to ensure that the object lies in

the 7th quadrant. When the object moves away from
7thquadrant, suitable signals are sent to the robot to take
action accordingly. Robot movement is given according
to the direction shown in the figure 5. The arrows depict
the movement of the robot based on the position of the
object.
 International Journal of Engineering Trends and Technology (IJETT) Volume 35 Number 4- May 2016

Quadrantbased algorithm is used to achieve the entire

task of picking and placing the object into its respective
coloured container. Figure 6 shows the conditions for the
robot movement.

5.2 MANUAL

The commands are given to the robot through an

application to perform the desired task.

Fig 9: Snapshot of image processing

7. FUTURE ENHANCEMENT

Tracking of multiple objects.

Position of container can be autonomous.
Area of coverage can be expanded.

Fig 7: Frontend of App 8. ACKNOWLEDGEMENT

Controlling the movement of the robot can be done using We would like to thank VGST (Vision Group on Science
some kind of remote control interface. It may be web and Technology), Government of Karnataka, India for
interface or a mobile interface. Since setting up python providing infrastructure facilities through the K-FIST
and SSH in raspberry pi is very easy without any cons, it Level II project at KSIT, CSE R&D Department,
is recommended to use a mobile based user interface for Bengaluru
controlling the robot motion. One of the most common
approach which is used to send command to the server is 9. REFERENCES
AppInventor.Frontend of the App which is shown
consists of a slot to enter the IP address of the sever to
connect to the raspberry pi. Various controlling buttons Detection and Path Find
IEEE.
are used to send the signals to the server and that will [2]Hyung-Bok Kim and Kwee-Bo Sim, A Particular Object
execute the corresponding controlling action. Tracking In An Environment Of Multiple Moving Objects
International Conference on Control,Automation and Systems
6. RESULT 2010.
[3]MaimunHujaHusin, FauzilianaOs
At the end of the automatic mode, robot tracks, analyses shape pattern recognition for FPGA-based object tracking
the colour of the picked object and drops the object into .
the respective coloured container. [4] Ana Maria Ocana,Francisco

At the end of manual mode, robot moves and does the Inc. 2013.
task as desired by the user according to the commands [5] https://en.Wikipedia.org/wiki/Green%27s_theorem
given through the application(app). [6] Jaya P and GeethuBalakrishnan, Contour Based Object
Tracking, IJCSIT 2014.
[7] Narayan Pandharinath Pawar & Minakshee M. Patil,
Driver Assistance System based on Raspberry Pi, International
Journal of Computer Applications (0975 8887) Volume 95
No.16, June 2014, pp. 36-39.