ml-cnn-for-crowd-management

Project For DevJam Team - Infiltrators

Table Of Contents

• Info
• Requirements
• Overview of code
• How To Run
• Timeline

Info

To ensure Social Distancing is being followed , We have two approaches:-

1. Queue Management in - front of retail shops. 2.Crowd counting and liminting the no. of people at a time the stores.

If either of the two is being unfollowed , we plan to send a message to the shop owner / manager.

This is the whole idea around which around project revovles.

Requirements

This project is being developed using

• Python 3.6
• OpenCV
• imutils
• Numpy

Overview

1. Queue Management

We implemented a tracking + detection system for controlling the Queue and Crowd outside the store.

For detection we are using the pre-built model present in OpenCV: hog = cv2.HOGDescriptor() hog.setSVMDetector(cv2.HOGDescriptor_getDefaultPeopleDetector())

The tracking algorithm consists of Centroid POint tracking And ID Managements. The centroidtracker.py file in OUTSIDE_STORE has all the further details.

There is a video also in OUTSIDE_STORE which is used as a demo for the basic working of the aforementioned scripts. The full functionality of this is still a Work In Progress.

2. Crowd Counting

How-to-run

Directory_where_the_project_is>cd OUTSIDE_STORE
>python queue.py

Timeline

17 - 05 - 2020

Crowd checking is being done thru the files in the OUTSIDE_STORE.
For testing we have used the video also included in the folder.
Position Markers - Work In Progress(Not Reacting to the number of people in frame)
So on temporary basis , have included some custom markers.
END.