IoT based User Behavior and Profiling for Super Store

Marts are one of the widely visited place in any region of the world. People irrespective of their age and gender, prefer going to marts rather than using online shopping mode because of this there is a dense crowd in the marts and management of mart space and workforce is becoming challenging. Inef

Project Title

IoT based User Behavior and Profiling for Super Store

Project Area of Specialization

Internet of Things

Project Summary

Marts are one of the widely visited place in any region of the world. People irrespective of their age and gender, prefer going to marts rather than using online shopping mode because of this there is a dense crowd in the marts and management of mart space and workforce is becoming challenging. Inefficient management of space, orientation and labor may result in monetary loss and losing customer, difficulty in branding and pitching new products.

Therefore, when new technologies offer various solutions to different application areas, mart management can also get the benefit. The proposed system acquires images using depth camera and then can count the people entering or leaving the vicinity by using computer vision. This will not only be beneficial from the security point of view but will also help to determine the point of attraction of the customers using the amount of time they spent at particular spot.

In this project, an IoT based framework is proposed as we are using wireless depth cameras at different points, it will use an image captured from the cameras to count the number of people in the mart. Firstly, a model is to be trained for object detection and then a dataset will be created that will store all the information. Secondly, the point of attraction of customers will be evaluated by checking the time they spent on a product/section. This can be achieved by checking the frame rate, if a person stays in the frame for a particular time then we can conclude that he/she has interest in that product and most of the customers are visiting at that section.

This information will be used for efficient placement of a product on the most visited spot, labor deployment in most crowded hours and days.

Project Objectives

To develop an IoT based user profiling and behavior system for super store that aims to manage the crowd by determining their behavior. This can be achieved by designing and implementing an efficient system that contains the following different parts:

• Counting of people entering or leaving the mart.
• Determining the point of attraction of the customers by checking the frame rate.
• Minimization of incidents and efficient utilization of resources in the mart.

Project Implementation Method

Our system will consist of the following features.

1. Firstly, we will determine the most efficient way to count the number of people entering or leaving the store using different algorithms and different depth cameras. After selecting the most suitable depth camera we will check for different angles for greater accuracy. The counting will be carried out through either head, body or face detection.
2. Secondly, we will determine the point of attraction of the customers using lesser number of cameras which will work on frame segmentation technique and then checking the number of frames to determine the time. If the customer is interested in a particular item, he/she will spend more time on that section and we can determine this by frame rate.

We are using datasets to train our system.

Benefits of the Project

• Marketing agencies can deploy this system.
• Resources can be utilized in a better way to improve sales.
• Minimization of incidents.
• Management of staff can be done in an efficient way.
• Anyone can use this project according to their need.

Technical Details of Final Deliverable

This project can be deployed in Super stores as well as in any building or any area in which we need to count people.

To count people entering or leaving, a camera is to be deployed in that area but camera should be over the head. The camera is programmed in the following way:

• It capture frames every second and a reference line is drawn, If a person crosses that line then the camera will decide whether the person is entering or leaving. A condition is applied in code that will decide, If a person is entering then the Entering count will be incremented and If a person is leaving then the decrementing count will increment. 

To determine Point of Attraction a camera is to be deployed in any building or super store. For super store it should be deployed in such a way that it captures more than one shelf. The camera is programmed in the following way:

• The camera will capture frames and then the frames are segmented and a region of interest is selected. If a person stays at any specific shelf then the camera will check If the person stays at that specific shelf for a particular/defined time then It'll count that person. Same procedure is for second shelf. In this way we can compare the count of people on first shelf with the count of people in second shelf and based on their count we can conclude that more people are interested in that area. 

Super stores can manage their resources by evaluating Point of Attraction.

Final Deliverable of the Project

HW/SW integrated system

Core Industry

IT

Other Industries

Core Technology

Internet of Things (IoT)

Other Technologies

Artificial Intelligence(AI)

Sustainable Development Goals

Industry, Innovation and Infrastructure

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com