Huffman-Clowers labelling using fuzzy logic

• Main Authors: Co, Brian O., Huang, Chung Liang, Patetico, Kendrick O., Yap, Christopher Rey C.
• Format: text
• Language: English
• Published: Animo Repository 1994
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/16596
• Institution: De La Salle University

Huffman-Clower labelling is one of the basic foundations in the development of robotic vision. Fuzzy logic and fuzzy set theory on the other hand has been extensively used for control purposes both in hardware and software applications. The primary objective of the study is to identify the perspective of 3-D objects from 2-D grabbed video images. The object is identified primarily through edge detection and its perspective known by identifying and classifying the individual lines making up that object according to the convention proposed by Huffman and Clowes. The Huffman-Clowes algorithm labes lines as either convex, concave or border lines. However, before these lines can be labelled properly, several image processing and analysis techniques had to be employed. This included the study of what dilation, erosion, and thinning are and how they worked. It also entailed the creation of a line detection function which could take on the rigors of sensitive line identification. Another challenging aspect of the study was incorporating fuzzy set theory and fuzzy logic into the labelling process. The Huffman-Clowes algorithm is simply not appropriate as an optimum fuzzy logic-based application. Still, the group managed to come up with a system of utilizing fuzzy set theory in an expert system environment.