DAI_final.ppt Social surveillance monitor

Social surveillance monitor

1. By:
Rutuja more(37)
Akash sirsat(51)
Narayan Fulmarkar(19)
Tukaram Shinde(4)9
Rohit Kulkarni(33)
Social Surveillance
monitor and crowd
controller

2. Index
Aim
Objective
Problem Statement
Scope of project
Feasibility study
Literature review
Front End and Back End
Algorithms used
Proposed Diagram
Conclusion

3. AIM
To build Social Surveillance monitor and crowd controller at mall/super markets
Monitor the crowd Ensure Limited crowed enters mall/supermarkets and no social
distancing violations occurs.
To decrease the spread of contagious diseases(Corona)
To check whether people are following the social distancing rule enforced by the
government.
To detect distances between people in a provided video.
To break the chain of infection of corona like diseases.

4. Objective
Social Surveillance monitor and crowed controller uses YOLO
Object detection algorithm to detect object such as people in
The public space (malls)
It should mark them as a object with the box
It should be able to differentiate people from other objects within the frame.
It should be able to detect violation of social distancing among people.
It must detect number of people to control crowd

5. Problem Statement
Monitor and control crowed at public places
Limited crowd enters public places
Social distancing is a method used to control the spread of contagious diseases.
To implement social distancing rule in practical we can use this project.
This project might help to decreases viral contacts among people
We want our application to provide uninterrupted continuous service throughout the
process.
It should not be crashed while social distance detection process

6. Scope of the Project
This project detects multiple objects within given frame.
It differentiate object as a person from other objects.
It calculates distance between objects and detects whether there has been
violation of social distancing.
It doesn’t necessarily delivers accurate output all the time.
It detects number of people within the surveillance or video

7. Feasibility of Project
Feasibility analysis determines the project’s potential for success
This project is technically feasible because it’s hardware requirements are
minimal.
This project delivers easy to interpret output to end users so we can say that it’s
operationally feasible as well.

8. Literature Survey
Person detection for social distancing and safety violation alert based
on segmented ROI.
Research paper based on Person detection for social distancing was published
recently in August 2020 by Malaysia.
It is basically a social distancing detector system that detects violation of social
distancing.
The system is developed by using Python 3, OpenCV and Caffe Framework.
MobileNet SSD model of Caffe framework is used for object detection to
determine the object as person.
Masking technique is used to estimate area of ROI.
The person’s bounding box is determined by comparing ground plane point with
ROI.
Then the center of bounding box is estimated and distance between multiples
centers are calculated to detect violation.
D:\TYMCA\EDI\Person_Detection_for_Social_Distancing_and_Safety_Violation_
Alert_based_on_Segmented_ROI.pdf

9. Literature Survey
Social Distancing Detection with deep learning model
Research paper based on Social distancing detection with deep learning was
published in 2020.
The system is developed using Python 3, openCV, YOLO v3.
The system has same purpose as that of previous one i.e. to detect social
distancing violation.
YOLO v3 is an algorithm used in this nomenclature for object detection.
Later on the centroids of bounding boxes of these objects are determined.
After determining centroids, distances between multiple centroids is calculated
and measured against required minimum distance norm.
If the distance between centroids is less than required distance then red
bounding box will be generated around objects to denote the violation.
D:\TYMCA\EDI\Social_Distancing_Detection_with_Deep_Learning_Model.pdf

10. Front End and Back End
AVAILABLE TECHNOLOGIES:
•Library: opencv, numpy
•Development platform: Pycharm
•Language used : Python
TOOLS USED:
•Editor used: Pycharm with Python
•Operating system: Windows 10
HARDWARE USED:
•Processor: Intel Core i5/i3
•RAM: 8 GB
•Camera
Proposed Algorithms:
•YOLO object detection algorithm
•Euclidean Distance Algorithm

11. Algorithm used
Yolo algorithmis used in this project to detect the objects in real time.
Yolo algorithm works by using following three techniques:
1] Residual blocks
Sample vector : [Pc,Bx,By,Bw,Bh,c1,c2]
2] Bounding box
3] Intersection over union

12. Algorithm used

13. Algorithm used
Euclidean distance formula is the one we’ve used in this project to calculate
pairwise distances between centroids of objects
Euclidean distance formula gives the distance between two points or it gives
a straight line distance.
 https://www.cuemath.com/euclidean-distance-formula/

14. Proposed Diagram

15. Conclusion
The project will monitor crowd and ensure limited crowd enters public place
no social distancing violations occurs.
The project will help to reduce the spread of contagious diseases like corona.
It will detect violation of social distancing parameter and enforce people to
follow the rules.
It will make people more aware regarding social distancing.

16. 16