Jaquar Interactive Floor

    Client: Jaquar
    Roles and Responsibilites:
  • • Complete Application Development
  • • Project Management
  • • On-Site Execution

Jaquar is a leading manufacturer of premium bath and kitchen fittings in India. For their experience center in Manesar, they commissioned another company to create an interactive floor projection in a huge area of 30 feet by 12 feet area, and 3 circular inlaid projection tables. They failed to make either of them work and backed out at the last moment. Digital Jalebi was then approached to create these 2 installations from scratch in one month.

The interactive floor installation is the first thing that the visitors experience just as they enter. It simulates the feeling of a water on road and walking through it forms ripples wherever they step. To achieve this, 4 Kinect v1 were placed around the area on the floor. Using their depth feed, I could detect any object on the floor. I made an openframeworks app that after laboriously calibrating all the Kinects to a single coordinate system, could pinpoint the location of any object on the floor.

Detecting user foot step was more complex than detecting position as we needed accuracy to mm scale. To do this, openCV was used to generate blobs of feet and using the lowest point on the centroid, and some velocity prediction, I could detect the time and point of contact. But still due to the delay induced in the system from capture cards and Dataton servers, the ripples would be slightly delayed(about 1-2 seconds) till they were displayed on the projector.

The interactive touch table are circular rear-projected glass that are inlaid fllush with the top surface into a corian table. It isn't possible to use standard touch screen on overlay as they are rectangular and would be visible over the table surface. So I decided to use the Rear DI(Diffused Illumination) method of creating touch-tables.

IR lamps illuminated the surface from below and a webcam withIR bandpass filter would detect and IR light reflected by user's touch. Since the finger reflects light even when hovering above the surface, careful calibration was required to get the touch only when light intensity was above a certain threshold and the galss was touched.

This video shows one of the three multi-touch surfaces in use. The application is a website running in chrome developed by another company.

This is a test video of user position detection for a single point. Circular projected markers highlight any object touching the surface of the floor.