Posted on Wed, March 19. The project must be submitted to the Teaching Assistant by Thursday April  03 (11:59pm). The project must be submitted by electronic mail to guerra@cs.umd.edu with a subject "CMSC 733: P2". You must submit only a readme file (readme.txt), your source files (Matlab .m), and, possibily, the input converted to another standard. Again, you are required to have three files named "A.m" and "B.m" among your source files. These files implement each part of the project, respectively. They must read the given input and display the expected output. You must tar your multiple files into a single file named p2.tar.

Overview.

The goal of this project is (A) to find the motion (rotation and translation) between any two stereo images in a video and (B) to detect independent movement in human motion video. You must use the Image Processing Toolbox of Matlab to implement your project.

Part A.

Step 0. Perform the calibration of the camera. Given a set of images (calib?.tif) of a calibration device, use the  Camera Calibration Toolbox  from CalTech in order to compute the calibration matrix.

Step 1. Compute stereo matchings in a pair of images (seq.2.????.tif ). Solve the feature correspondence problem for the given images using Project 1 - Part A.

Step 2. Find the Fundamental Matrix. Given the matches from Step 1, built a linear system and solve for the Fundamental Matrix. You may use the techniques described in the text book (e.g. eight-point algorithm), any improved version of them (e.g. normalized eight-point algorithm), or a statistically robust approach as RANSAC.

Step 3. Compute the Essential Matrix. Given the calibration and the fundamental matrices, the Essential Matrix is trivially defined.

Step 4. Solve the Realtive Orientation problem. Given the Essential Matrix, find the rotation and the tranlation of the associated rigid motion. You may use the algorithm described in the textbook.

In part A, the expected output is the first image (alphanumerical order) with the disparity vectors for the point matches used and the corresponding epipolar lines.

Part B.

The independent motion detection problem consists in detecting moving objects by an independent moving observer. You should review the related literature and use what you learned in class to design and implement an algorithm that solves this problem.

The input for Part B is a sequence (seq1.x.????.tif ) acquired from four cameras (x = 0,1,2,3). You should use only 0 and 1. The sequence consists of a video with independent motion. Calibration information is also provided in the video cal1.x.????.tif.

The output for Part B is the first image of video seq1.1.????.tif with only the detected moving object being shown. Note that the background should be erased. You may also provide a complete video only with the moving object.

Submission Files.

You must submit a readme file, named readme.txt. The readme file enumerates and describes all files in you tar file. The readme file should also contain a description of your implementations. You must also submit your source (Matlab .m) files. All multiple files must be in a single tar file (p2.tar) and sent to guerra@cs.umd.edu.