Optimal Polygonal Approximation of Digitized Curves Using a Genetic Algorithm Combined with Perimeter Object Function

• Main Authors: Kuo-lung Chu, 朱國隆
• Other Authors: Ching-yu Yang
• Format: Others
• Language: zh-TW
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/33513246258524410256

碩士 === 國立高雄應用科技大學 === 模具工程系碩士班 === 92 === Polygonal approximation divides into two catalogers, one is optimal polygonal approximation that uses the optimization algorithms to solve the problem, and the other is the heuristic algorithms that use the certain operational rule to approximate the polygonal. In this paper, testify of the polygonal approximation can be divided two part. One is the performance of the perturbations of data that pretend perturbations of data by shift or rotation the data when capture image in instability environment in order to test polygonal approximation algorithms. The other changes the algorithms scale parameters that pretend perturbations of data by change scale parameter when capture image in different distant to compare the polygonal approximation algorithms. The results show that genetic algorithm combined with maximum perimeter of object function is stable and reliable whatever the image picture is rotated or shifted. The chromosome of perimeter is the same value in image picture whatever the image picture is shifted or rotated. Therefore, genetic algorithms combined with maximum perimeter objection function that performs well in most of cases.