3D Room Mapping with Handheld Time of Flight-Camera
Tell Me More

Project Description

The goal of this project has been to develop a program that can produce a 3D point cloud of a room from a sequence of Time-of-Flight data. The data has been collected using a Fotonic G-series camera which produces depth images. There has been three main parts in this project; point-cloud logging from sensor, registration and outlier filtering of the point clouds. On top of this, a mesh of the resulting model has been used to visualize the results.

Hardware

The hardware used in this project is a Fotonic G-series Time-of-Flight camera.

Course

The project is part of the course: TSBB11 Images and Graphics, Project Course CDIO, at Linköping University.

Client

The client for this project is Per-Erik Forssén, associate professor at the Computer Vision Laboratory at Linköping University, in association with Fotonic.

Software

The software captures several point clouds from a Fotonic G-series depth camera and process them to make a global coordinate system and output a mesh. The software is implemented using C++ and the Point Cloud Library, registration is done using Fast Global Registration.

Data Collection

The point cloud data comes from the camera in a three dimensional coordinate system.

Pre Processing

The data is pre-processed by down-sampling, filtering and normal estimation to be able to extract features from the point cloud.

Registration

For each pair of point clouds a transformation is calculated by the algorithm Fast Global Registration, which utilizes FPFH features.

Visualization

The finished registered point cloud and mesh is visualized using the Point Cloud Library.

Documents

The project contained several documents to put into writing the different tasks and aspects of the project.

Visualiser

The visualiser is used to help visualise 3D objects on a flat surface. Choose a new color by pressing c, increase dot size with + and decrease with -.



Team

The project team consisted of four members from the signal- and image processing master specialization at Linköping University

Anna Hjelmberg

Developer

Tobias Grundström

Developer

Mats Nilsson

Developer

Fredrik Olsson

Developer