EESC 7V85 – 501: Vision-Based Estimation and Control
Fall 2012, University of Texas at Dallas
This page is under construction and subject to change
Fall Semester 2012
Tuesday and Thursday 1:00PM - 2:15 PM
|Office Location||ECSN 4.910|
Syllabus - EE_7V85-Syllabus_F_2012_Gans.pdf (please note that the syllabus is preliminary and subject to change until the start of the course)
EESC 7V85 – 501 Vision-Based Estimation and Control (3 semester hours) This is a course for graduate students in ECE, CS, ME, Math and Geosciences. Topics cover building 3-D graphical models from 2-D images from multiple or single cameras, recovering 3-D motion of mobile robots from onboard cameras. Specific topics include:
Feature detection and tracking
Structure from motion
Optical flow-based estimation
Position Based visual servoing
Image-Based visual servoing
The class is self-contained and will provide introductions to robot kinematics, linear and nonlinear control and Kalman Filtering.
There are no planned exams. Course work and assignments will be heavily project oriented, with programming assignments to implement class topics in simulations and real images/video. By the end of the course, students will have a valuable set of functions and programs to use for future work or research. This will culminate with a final project to be chosen by the student in conjunction with his or her research advisor, if desired. This project is intended to be of the level that it could lead to a conference paper, and students will present their project to the class.
Pre-requisites& other restrictions
There are no formal prerequisite courses, but you do need the knowledge/skills listed below. You should contact me prior to enrolling.
This course requires a solid background in linear algebra such as matrix representation, linear dependence, null spaces, eigenvalues, eigenvectors, and singular-value decomposition. At UTD, this would be found in EE2300 or EE6331.
There is no prior knowledge in computer vision, robotics or control theory required. This course may be taken independently of or together with the other computer vision, robotics or control theory courses. Previous experience with these topics may be of benefit.
Familiarity with Matlab is highly recommended. If you are unfamiliar check out http://www.math.ucsd.edu/~bdriver/21d-s99/matlab-primer.html.
Texts & Materials
Highly Recommended: An Invitation to 3-D Vision: From Images to Geometric Models by Y Ma, S. Soatto, J. Kosecka, and S. Sastry
Supplementary: Introductory Techniques for 3-D Computer Vision by E. Trucco and A. Verri
Multiple View Geometry in Computer Vision R. Hartley and A. Zisserman
Other materials, such as journal papers, will be provided in class