基于全局和声搜索算法的椭圆拟合  

FEllipse fitting based on global harmony search algorithm

作　　者：雍龙泉[1] 张媛媛 黎延海[1] YONG Longquan;ZHANG Yuanyuan;LI Yanhai(School of Mathematics and Computer Science,Shaanxi Key Laboratory of Industrial Automation,Shaanxi University of Technology,Hanzhong 723001,China)

机构地区：[1]陕西理工大学数学与计算机科学学院,陕西省工业自动化重点实验室,陕西汉中723001

出　　处：《安徽大学学报(自然科学版)》2025年第1期1-7,共7页Journal of Anhui University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(11401357);陕西省自然科学基础研究计划项目(2024JC-YBMS-014);陕西省教育厅青年创新团队项目(23JP024);陕西理工大学科研项目(SLGNL202409)。

摘　　要：建立了椭圆拟合问题的约束优化模型,利用绝对值函数给出了一种约束处理方法,将原问题转化为无约束优化,采用全局和声搜索算法求解.数值实验分别对长轴和短轴在坐标轴上、长轴和短轴不在坐标轴上的椭圆拟合问题进行了研究,结果表明在数据没有异常值的条件下,即使有噪声,拟合结果也较好.A constrained optimization model for ellipse fitting problem was established,and a constraint processing method was presented by using absolute value function.The original problem was transformed into unconstrained optimization and solved by global harmony search algorithm.Numerical experiments were carried out to study the ellipse fitting problem of long axis and short axis on coordinate axis and long axis and short axis not on coordinate axis respectively.The results showed that the fitting results were good even if there was noise in the data without outliers.

关 键 词：椭圆拟合 绝对值函数 约束优化 全局和声搜索算法 

分 类 号：O212[理学—概率论与数理统计] TP18[理学—数学]