Wang, Ching-Cheng (Professor of industrial engineering)
M.S. (Master of Science)
Department of Industrial Engineering
Form perception; Computer vision
Machine vision employs automatic interpretation for control purposes such as process control, quality control and robot control. Automated object recognition is necessary for the aforementioned tasks in an intelligent manufacturing environment, one of the most important frontiers of machine vision research. The most widely investigated techniques for object recognition have traditionally been pattern recognition methods using monoview or stereo images. These techniques have enjoyed wide-spread use. However, they are viewpoint dependent. Thus, the reliability of these methods is typically poor unless the to-be-recognized object is placed with a narrow position inaccuracy. Unfortunately, it is difficult to handle object without errors being made. Therefore, the viewpoint independent object recognition methods have been explored using characteristics o f conditional and absolute invariance under the perspective transformation. Recently, Lei proposed a viewpoint independent method which adopts the cross ratio as the shape register. But since uncertainties exist in visual mensurations and part dimensions, the successful application of Lei’s method in distinguishing pooled parts requires the calculation of the shape register’s range when subject to mensuration and dimensional errors. This work explores the formula needed for calculating the range of the viewpoint invariant shape register.
Chitnis, Abhijeet, "Viewpoint invariant automated object recognition" (1992). Graduate Research Theses & Dissertations. 6650.
vii, 65 pages
Northern Illinois University
Rights Statement 2
NIU theses are protected by copyright. They may be viewed from Huskie Commons for any purpose, but reproduction or distribution in any format is prohibited without the written permission of the authors.