• rbdlab haifa

Personal Tour Guide

Agenda

  • What? Why? How? (3D computer vision – triangulation, people detector, OptiTrac, Ladybird)

  • Pre-knowledge (Matlab, OpenCV, Odroid, AWS, Arduino, Network Communication)

  • First Approach ( People detector in openCV, Calibration, Stereo Calibration, Triangulation)

  • The important milestone (Matlab detector and triangulation)

  • The final Milestone – OpenCV, Odroid, AWS, N-cameras people detector)

  • Summary

Vision & Goals

Indoor navigation using cameras.

  • Let’s divide today’s world navigation to indoor and outdoor environments.

  • Outdoor: There are several solutions that the best known is the GPS

  • Indoor: There are some solutions out there which are mainly based on WiFI / Bluetooth / etc.


  • All indoor solutions today are not robust and not widely used.

  • We are trying to suggest a different approach using only computer vision power with home low cost equipment.


  • People Detector in ‘Cartesian coordinate system’ – We want to detect human body in the relative to the real world coordinates system.


Prerequisites

Stereo Vision

  • What is stereo vision? Each eye captures its own view and the two separate images are sent on to the brain for processing. The mind combines the two images. It is a three-dimensional stereo picture.