ECE 472/561 Robotics and Vision
1. Devise a wire frame representation of a house. Specifically, the house is represented by 10 or more vertices in world coordinates. Simulate a camera and compute the pixel coordinates for each vertices. Draw the house using these pixel coordinates and matlab’s 2D plotting functions (e.g. plot, line).
2. Rotate the camera around the house to 4 different views and redraw its image.
3. Construct a virtual calibration object (e.g. checkerboard cube). Simulate the image of a calibration object and use ginput to detect the locations of its vertices. Use the DLT (direct linear transform method) to recover the calibration matrix. Compare the known matrix to the recovered calibration matrix.
4. Reproject the points onto the virtual calibration object so that the original calibration object can be seen in the same frame as the reprojected points. Comment on the error (difference between the original and reprojected points).
5. For Graduate Students Find two images of the same plane (billboard, floor, etc). Find corresponding points using ginput. Find the homography using the corresponding points. Warp the planar image regions into alignment.