Team Name: TextureIt!

Team Members: Austin Reiter and Hao Dang

Download the final report.

1. I.Project Description -
   

        Build a 3D scanner capable of dense reconstruction of small objects.

Need to overcome the typical challenges of using cameras:   

    - correspondence problem

    - how to deal with textureless regions?

    - matching features from different viewpoints

2. II.Methods -
   

    A. Photometric Stereo

        - typically used to recover surface normals by varying the lighting conditions

        - can we also recover 3D geometry (i.e., point clouds)?

        - attach a light source to a robot arm and move to known positions

        - image the object using a high-dynamic range camera

    B. Two-camera stereo system with a laser pointer

        - use a laser pointer to avoid having to match any features

        - control the laser with a robot arm for fine movements and dense coverage of the object

III. Hardware -

    2 Prosilica Gigabit Ethernet cameras with 25 mm lenses (image resolution of 1620x1220 pixels)

    1 Point Grey Dragonly2 with two modes:

        (1) High-dynamic range with 12 bits per pixel (image resolution 1024x768 pixels)

        (2) RGB Color mode, 8 bits per pixel (image resolution 1024x768 pixels)

    1 Simple laser pointer

    1 Staubli Robot Arm (programmable through C++)

IV. Results -

    Scanning a small bird statue with method II-B:

                            Figure: (top-left) The texture color image of the bird statue, used for texture mapping the mesh.  (top-right) The textured, triangulated mesh after reconstruction from the stereo system.  (bottom) The raw point cloud obtained from the stereo system.  These points are used to create a triangulated surface mesh which is then texture-mapped for photorealistic imagery.