Moving object detection in video surveillance

Heaven’s Light is Our Guide
Rajshahi University Of Engineering & Technology
Department Of Electronics & Telecommunication Engineering
Moving Object Detection in Video Surveillance
PRESENTED BY
Md. Ashfaqul Haque
Roll: 124006
Md. Sharful Insan
Roll: 124007
Department of Electronics &
Telecommunication Engineering,
RUET
SUPERVISED BY
Prof. Dr. Md. Rabiul Islam
Head,
Department of Computer
Science & Engineering, RUET
1

Outlines
 Introduction
 Motivation
 Objectives
 Related Works
 Proposed Methodology
 Flowchart
 Workflow
 Experimental Results
 Conclusions
 Limitations & Future Work
 References
2

Introduction
 Object Detection is a computer
technology related to computer
vision.
 Detection is basically locating
object in an image or a video
sequence.
 Moving object detection
applications are car detection,
person identification or wild life
monitoring.
 It is used in security monitoring,
intelligent perception . Figure[9]: Moving Object Detection
3

Motivation
 Object detection system introduces a new technology
of security.
 It is used different organization, institution, offices and
social sites.
 Object Detection is used to control and reduce ,
terrorism, crime, robbery, shop lifting, and accidents.
4

Objectives
 Develop a computational model to identify the moving
objects by using background subtraction
 Detect the moving objects in various scenarios
 Develop a comparative result of efficiency for better
object detection
 Develop an application for the smart surveillance system
using object detection
5

Related Works
A great deal of work have been done on object
detection using various method
 An approach is used for moving object detection and
tracking in indoor environment[1].
 A Detection method of moving object based on
background subtraction[2].
 An efficient real time moving object detection method for
video surveillance[3].
 A moving objects detection algorithm based on improved
background subtraction[4].
 Motion human detection based on background
Subtraction[5].
6

Related Works(cont.)
 Moving object detection in spatial domain using
background removal techniques - State-of-Art[6].
 A method of detecting and tracking object in wide
area surveillance using Thermal Imagery[7].
 A method based on the kernel method is used to
detect and tracking moving object[8].
7

Proposed Methodology
 In our proposed method we have used Background
Subtraction.
 First we have taken frames from the video sequences
and extracted the background frame and the current
frame. Then subtract the background frame from the
current frame.
 Then we have used some filtering techniques to
remove noisy areas.
 Finally we have used morphological operation on the
subtraction frame and get the final output.
8

Flowchart
Input Video
Sequences
Extract Frame
Convert to Gray
Scale
Convert to HSV
Moving Frame &
Background
Frame
Output
Morphological
operation
Boundary
Labelling
Binary Image
Median Filtering
Background
Subtraction
9

Workflow
 Take a video sequence
 From the sequence extract
the background and moving
frame
Figure: Background Frame
Figure : Moving Frame
10

Workflow (cont.)
 Convert the extracted frame
to the HSV format
 H – Hue, S – Saturation,
V - Value
 Hue represents the color type
 Saturation represents the
vibrancy of the color
 Value represents the
brightness of the color
Figure : HSV image
11

Workflow(cont.)
 Subtract the background frame from the moving frame
and get subtracted image. Here subtraction is
performed using the bitwise XOR
 A bitwise XOR takes two bit patterns of equal length
and performs the logical exclusive OR operation on
each pair of corresponding bits
 Bitwise XOR operation
0 XOR 0 = 0
0 XOR 1 = 1
1 XOR 0 = 1
1 XOR 1 = 0
12

Workflow(cont.)
 Then convert image to
gray scale
Figure : Gray scale image
 Gray scale image is
converted to the Binary
image
 Luminance greater than
level with the value 1
(white) and all other pixels
with the value 0 (black)
Figure : Binary image
13

Workflow(cont.)
 Here the Binary image is
filtered using median
filtering to remove noises
Figure : Median filtered image
 Boundary labeling have
been used to remove the
areas which are less
connected or not
connected
Figure : Boundary labelling
14

Workflow(cont.)
 Closing is defined simply as
a dilation followed by an
erosion using the same
structuring element for both
operations
 Closing operation have
been used to remove the
holes inside the image
objects Figure : Image after closing
15

Experimental Results
 We have implemented our method for detecting
object using MATLAB software
 We have tested our implemented method by the video
captured by camera and sequences collected from
internet. The sequences are captured in different
scenarios
 We have calculated the accuracy for our proposed
method and compare with the other existing method
16

Experimental Results(cont.)
 Moving object detection result
for different indoor scenarios
17

 Moving object detection in
various outdoor scenarios
18

Accuracy Calculation :
 Calculate the true positive(TP), true negative(TN), false
negative(FN) and false positive(FP) value comparing
with ground truth value to measure accuracy
 TP : Detected value matched with the truth value
 TN : No value both in the detection and ground truth
 FN : Rejected in detection but present in the ground
truth
 FP : Wrong detection but not present in the ground
truth
19

Accuracy Calculation(cont.) :
 Calculation of accuracy using the TP, TN, FN, FP
Accuracy =(
(TP+TN)
(TP+TN+FN+FP)
×100)%
20

Accuracy Calculation(cont.) :
 Accuracy for various scenarios :Scenarios Accuracy (%)
Camouflage 94.9898
Rainy 94.9606
Indoor 93.8982
Bootstrap 94.5482
Office 93.5676
Sunny 93.9579
 The average accuracy of our procedure is 94.3204 %
21

 Accuracy comparison with
Kim & Hawng method and
Dewan & Chae method
Method Detection
Rate (%)
Kim & Hawng Method 91.02
Dewan & Chae Method 91.07
Proposed Method 94.3204
22

ROC Curve Generation
 True Positive Rate (TPR)
TPR=
TP
TP+FN
 False Positive Rate(FPR)
FPR=1-
TN
FP+TN
Experimental Results(cont.) 23

 ROC curve for our
proposed method
24

Conclusion
 A background subtraction technique for detecting
object has been proposed here
 We have got accuracy for the proposed method
94.3204 % which is better than the Kim & Hwang
method and Dewan & Chae method
 Moving object detection is always a challenging task
and there is a room to improve this method
25

Limitations & Future Work
 Our method can not detect the moving target in heavy
rain condition and other natural calamities
 It can not detect moving target in long distance video
sequences
 We have to reduce this limitations in future
26

References
[1] Shucai Wang, Liying Su, Kim Wang and Yueqing Yu, “Moving Object
Detection and Tracking in Indoor Environment”. 2011 International Conference
on Electronic & Mechanical Engineering and Information Technology.
[2] Mr. Mahesh C. Pawaskar1, Mr. N. S.Narkhede2 and Mr. Saurabh S. Athalye1,
“Detection Of Moving Object Based On Background Subtraction” Volume 3,
Issue 3, May-June 2014.
[3] Pranab Kumar Dhar*, Mohammad Ibrahim Khan*, Ashoke Kumar Sen Gupta ,
“An Efficient Real Time Moving Object Detection Method for Video
Surveillance System” International Journal of Signal Processing, Image
Processing and Pattern Recognition. Vol. 5, No. 3, September, 2012
[4] Niu Lianqiang and Nan Jiang, "A moving objects detection algorithm based
on improved background subtraction," Intelligent Systems Design and
Applications, 2008, ISDA '08, Eighth International Conference on Volume 3, 26-
28 Nov. 2008, pp 604 – 607.
[5] Lijing Zhang, Yingli Liang “Motion human detection based on background
Subtraction” 2010 Second International Workshop on Education Technology
and Computer Science, pp 284-287.
27

References
[6] Shireen Y. Elhabian, Khaled M. El-Sayed and Sumaya H. Ahmed, “Moving
Object Detection in Spatial Domain using Background Removal Techniques -
State-of-Art” Recent Patents on Computer Science 2008, 1, 32-54.
[7] Santosh Bhusal, “Object Detection and Tracking in Wide Area Surveillance
Using Thermal Imagery”.
[8] Huanhai Yang ,Shandong Institute of Business and Technology, Yantai,
Shandong 264005, China, “Research on the Detection and Tracking of Moving
Target based on Kernel Method” Vol.8, No.2 (2015),pp.91-100.
[9]
https://www.google.com/search?q=moving+object+detection&oq=movi
ng+&aqs=chrome.0.69i59j69i61j69i57j69i61l2j0.3548j0j7&sourceid=chro
me&ie=UTF-8
28

Moving object detection in video surveillance

More Related Content

What's hot (20)

Similar to Moving object detection in video surveillance (20)

More from Ashfaqul Haque John (7)

Recently uploaded (20)

Moving object detection in video surveillance