亚像素边缘检测算法综述  

Review of sub-pixel edge detection algorithms

作　　者：曾猛杰 汪晨曦 赖俊杰 陈奕涵 陈泽伟 颜丙功 任洪亮 ZENG Mengjie;WANG Chenxi;LAI Junjie;CHEN Yihan;CHEN Zewei;YAN Binggong;REN Hongliang(College of Information Science and Engineering,Fujian Key Laboratory of Light Propagation and Transformation,Huaqiao University,Xiamen 361021,China;College of Mechanical Engineering and Automation,Huaqiao University,Xiamen 361021,China)

机构地区：[1]华侨大学信息科学与工程学院福建省光传输与变换重点实验室,福建厦门361021 [2]华侨大学机电及自动化学院,福建厦门361021

出　　处：《光学精密工程》2024年第23期3513-3524,共12页Optics and Precision Engineering

基　　金：厦门市高校科研院所产学研项目(No.20231ZC016)。

摘　　要：综述了亚像素边缘检测算法的研究现状及其应用进展。首先简要介绍了三类传统边缘检测算法(基于梯度、基于统计特性和结构化边缘检测算法)的基本原理,并分析了它们在边缘检测精度方面的局限性。其次,阐述了亚像素边缘检测的基本概念及其在提升边缘检测精度方面的优势。随后重点对三种主要的亚像素边缘检测方法(插值法、拟合法和矩方法)的理论基础、算法原理及其应用特点进行了详细论述。通过对比分析表明,这些方法在不同应用场景下均能实现高精度的边缘检测。最后,总结了亚像素边缘检测技术目前存在的主要问题,并从噪声抑制、模型优化和技术融合等方面对未来研究方向进行了展望。This paper reviews the research status and application progress of sub-pixel edge detection algorithms.First,three traditional edge detection algorithms(gradient-based,statistical-based,and structured edge detection algorithms)are briefly introduced,and their limitations in edge detection accuracy are analyzed.Then,the concept of sub-pixel edge detection and its advantages in improving edge detection precision are elaborated.Subsequently,three major sub-pixel edge detection methods(interpolation method,fitting method,and moment method)are discussed in detail regarding their theoretical foundations,algorithmic principles,and application characteristics.Comparative analysis shows that these methods can achieve high-precision edge detection in different application scenarios.Finally,the paper summarizes the main existing challenges in sub-pixel edge detection technology and provides future research prospects from aspects of noise suppression,modal optimization,and technology integration.

关 键 词：亚像素边缘检测 CANNY算子 插值法 拟合法 矩方法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] O439[自动化与计算机技术—计算机科学与技术]