ECE 595 Computer Vision Assignment 4

Submit Matlab code, images, and a Word document describing your work.

1. Construct a cube model including your picture on one of the faces. The original class images are in an Isidore attachment. You will need to crop a square subimage for each cube face and you may wish to reduce the resolution.

2. Generate a photo montage showing various views of the cube. Include several views showing the effect of perspective projection with the camera angle varying from large (maybe 60-degrees to small 10-degrees).

3. Scan the faces of some boxed product and generate and accurate model of that box. Save the box for possible use later in the semester.

4. Given the scanned image of a checkboard target and a ruler (Isidore attachment), find the size of the checkers as accurately as possible. Verify whether or not the individual squares are the same size. Use an automated edge or corner measurment rather than manual methods.

5. Given the collection of images of the HP3 robot moving a calibration target, find the displacement of the target (set 1) and orientation of the target (set 2). Experiment with various implementations of edge and corner algorithms. The image sequences are attached in Isidore as MATLAB .mat files.

6. Examine the background for the images above, and use it to determine the fractional distortion of the camera.

7. Capture a set of images of the KL corridor wall assigned to you. Superimpose colored edges on grayscale versions of the images. Experiment with different edge algorithms. Identify features whose coordinates you can obtain from direct measurment (ground truth).

Maintained by John Loomis, last updated 27 Feb 2011