Croweder Predicting people's based on their heights

With COVID regulations, came restrictions for the number of people entering stores. These restrictions often resulted in confusion among customers on when they are allowed in or not. Sometimes, a store patron is required to count heads in the store.

The aim of this project is to count the number of people going inside of a store and triggering a light/locking the door when a limit has been reached. The counting of the number of people entering a store is achieved by using an ultrasonic sensor along with a PIR sensor, and counting depending on which sensor is triggered first. The store count along with height data are sent to a cloud server via MQTT that is parsed into a MySQL Db for later analysis on store traffic and forecasting. Grafana Dashboard is being used to measure the real time operations such as histogram of people’s height in a particular period, line charts of people’s height as well as the quantity of people travelling in and out.

The cloud server is a private infrastructure built specially for this project, using Debian Buster version 10, with Python 3.9, Grafana, and NodeJS installed on it. It runs automated tasks for forecasting values and feeding the MySQL database in order to be displayed on Grafana Dashboard.

A forecasting model was created that could use previous count data to make predictions on store traffic based on the time/day. An GUI was also created in the form of an end user dashboard that displays the current store count, and allows for changing the maximum person limit for the led to actuate.