Broomstick

Inspiration

Data set provided by the challenge shows a significant potential for an integrated building facility management app.

What it does

Basic facility management for building care-takers and users. Users will be able to observe free desks spaces and meeting room from the app as well as report any faults to the care taker by opening tickets. Fault report function allows user of the building facility to report any fault, with a picture from their camera. The system is also able to process video feeds from CCTVs and determine whether cleaning staff needs to be dispatched for cleaning.

How I built it

With Meteor.JS for the webapp, Python and OpenCV for image processing.

Challenges I ran into

Intelligent Building

Accomplishments that I'm proud of

An image/video processing system that is able to classify clean tables.

What I learned

Image processing, classification, data analytic, Meteor.JS programming,

What's next for Broomstick

If wider variety of data is available, we will be able to extend the application to cleaning schedule management, office and disposable item restoking plan and optimization. The elevator control API together with restaurant POS data can potentially be used for the lunch time rush hour scheduling but that is not possible with the limited available data.

Built With

Submitted to

Junction 2016

Created by

I worked on the frontend with Meteor.JS and bootstrap with user interface design.

Isabel Luk
I worked on the structure of the Meteor.JS project and was responsible for fetching data from the API provided. I made visualization with the information provided and created a system that is able to show desk occupancy in near-realtime.

Steven
I worked on the ticketing system. It is based on the Meteor Framework with an implementation of Googles Material Design to implement nicely into the mobile phone environment.
It's possible to access the camera of the device to make a picture of the specified area.
A picture says more than thousand words.

Thomas Krug
I worked on image processing with python and OpenCV.
The goal was to distinguish between a tidy and a messy desk and e.g. remind people to clean up when they leave a conference room.
The algorithm captures a video stream monitors a predefined region of interest (desk) and uses histogram based thresholding to segment background from foreground objects. If it detects some objects it labels the desk as messy.

bookdo