Human Entrance Counter with PeekingDuck

Human Entrance Counter is a jointly-collaborated project between SPAI and SP-RITE and its aim is to detect the amount of human traffic moving in and out of a bounded region at any given time.

Project Overview

The project utilizes PeekingDuck with custom nodes for human detection and tracking along with Express.js as backend controller.

Software Required:

• Anaconda
• git
• GitHub Desktop (For those that are unfamiliar with CLI)
• VS Code

Getting Started

1. Clone the Repository with
   • Git CLI

     git clone https://github.com/SPAI-Team/Human-Entrance-Counter.git
     

   • GitHub Desktop App
     1. Add repository
     2. Clone repository
     3. Input the URL SPAI-Team/Human-Entrance-Counter
2. Setting Up Conda Python Environment

   conda create --name human-entrance python=3.8
   conda activate human-entrance
   pip install peekingduck

3. Open the ./human_counter directory in your terminal by running

   cd human_counter

4. Run peekingduck through the terminal:

   peekingduck run

5. Stop Peekingduck by pressing ctrl+c in the terminal or q on the peekingduck window.

Backend API Endpoints and Formatting

column	format	data type	example
time	YYYYMMDDhhmmss	string	20211114102300
footfall	int	integer	1
location	locationname	string	fc3

Get Footfall History of a Location

GET /history/:location/:startTime/:endTime

• Inputs:

  • locationname(string)
  • startTime(string in YYYYMMDDhhmmss)
  • endTime(string in YYYYMMDDhhmmss)

• Outputs:

  • json array of records with each records having:
    • footfallid (int)
    • time (string in YYYYMMDDhhmmss)
    • curentfootfall (int)
    • location (string)
    • netfootfall (int)

• Example: get /history/fc1/20211111024000/20211114130000

  output:

  [
      {
          "footfallid": 34,
          "time": "20211114101000",
          "currentfootfall": 20,
          "location": "fc1",
          "netfootfall": 20
      },
      {
          "footfallid": 35,
          "time": "20211114111000",
          "currentfootfall": 40,
          "location": "fc1",
          "netfootfall": 20
      }
  ]
  

  currentfootfall = total footfall(all added together), netfootfall = footfall recorded at time of record

Get Latest Footfall of a Location

GET /latest/:location

• Inputs:

  • locationname (string)

• Output:

  • json object with 1 record
    • footfallid (int)
    • time (string in YYYYMMDDhhmmss)
    • curentfootfall (int)
    • location (string)

• Example: get /latest/fc6

  output:

  {
      "footfallid": 33,
      "time": "20211111024500",
      "currentfootfall": 0,
      "location": "fc6"
  }
  

Get Latest Footfall of a Location within a set number of hours

GET /latest/:location/:noHours

• Inputs:

  • locationname (string)
  • noHours (string)

• Output:

  • json object with 1 record
    • footfallid (int)
    • time (string in YYYYMMDDhhmmss)
    • curentfootfall (int)
    • location (string)

• Example: get /latest/fc6/1

  output:

  {
      "footfallid": 33,
      "time": "20211111024500",
      "currentfootfall": 0,
      "location": "fc6"
  }
  

Record Footfall History

POST /history

• Inputs:

  • locationname(string)
  • startTime(string in YYYYMMDDhhmmss)
  • netfootfall (int)

• Output:

  • empty array, []

• Example: post /history

  body:

  time:20211114121500
  netfootfall:30
  location:fc1
  

  output: []