You are on page 1of 3

Analysis of various Artificial Intelligence techniques to monitor

social distance in crowd places


Abstract
Crowd management is a challenging scientific topic directly impacting on
the functioning of trafficked urban infrastructures such as, e.g., train, metro
stations, and markets to manage the crowd in the above-said places is a key
challenge for the government to control and maintain a minimal physical
distance (often also called “social distance”) between any two individuals. The
six feet distance needs to be maintained by the two people.
There are however several challenges associated with the automated
monitoring of physical distancing in crowds. First, we need to capture the image
by recording the content in the video and need to check whether the people
follow the distance in regular intervals of time. Then by using an algorithm we
need to check the values in a specific interval of time. If any people did not
follow the rule in the crowded place the message will be sent to the concerned
security person who is on duty there. He needs to warn them or put spot fine
based on the situation
Control room Camera1

o3
33
Camera3
Police or Security

o1
o4
!
o2
o5

Camera2 Police or
Security

. Camera4

Figure1: Placing the camera in various places


R1 20m
video 1
100 m road R2 20m

R3 20m
Video 2
R4 20m
R5 20m

Figure2: using video surveillance, we record the places


Steps for capturing the places along the camera
1. In 100 m Road we are placing 2 videos and splitting the road has 5
Regions
2. One video1 will cover the first 50m and another 50 m will cover by video
2
3. The video must be converted as a frame by frame and given as an input

I/P as frame
by fame

Analysing by
frame by
frame in each
seconds

Figure3: Analysing the video

4. Analysing the image as a region by region if we get the output in any


region value is more than 4 people for continuously then we decide its
crowd and sends the alert msg to control room
5. If suppose the value is less than 4 then we need to calculate each object
distance is weather less than 6ft. in this also we are getting less than 6 ft
then we decide crowd or more than 6 ft for each object then we decided
that there is no crowd
6. This analysis must be done for some interval of time
Table1: Analysis report
Time Value Decision
10:00 AM 4 No crowd
10:01 AM 4 No crowd
10:02 AM 5 crowd
10:03 AM 4 No crowd
10:04 AM 3 No crowd
10:05 AM 5 crowd
10:06 AM 6 crowd
10:07 AM 3 No crowd
10:08 AM 4 No crowd
10:09 AM 5 crowd
10:10 AM 3 No crowd

No crowd: 7
Crowd: 4
Final decision there is No crowd
Implementation:
Any ML algorithm by training the value we
make decision

You might also like