Undergraduate Computer Vision Curriculum To Complement A Robotics Program

Abstract

This article discusses a computer vision curriculum, including laboratory exercises, which is suitable for undergraduate engineering students. While classroom and laboratory exercises focus on off-line computation, on-line implementation can be achieved with simple equipment such as web-cams. Exercises include a sidewalk or line following exercise utilizing the Hough transform, a face recognition using eigenfaces, barcode reading, handwriting recognition, and sign language recognition. Data-set development for these exercises is also discussed. MATLAB and the Image Processing Toolbox are utilized to allow students to focus on higher-level understanding of commonly available image processing tools. The use of advanced tools allows students to attempt and finish meaningful examples. This paper focuses on exercises that serve as a useful complement to robotics curriculum and student robotics projects.