基于改进的变分水平集C-V图像分割模型研究  

Research on C-V Image Segmentation Model Based on Improved Variational Level Set

作　　者：赵银善 吐尔洪江·阿布都克力木[2] ZHAO Yin-shan;Turghunjan Abdukirim Turki(Xinjiang Agricultural Vocational Technical College,Changji 831100,China;College of Mathematics Sciences,Xinjiang Normal University,Urumqi 830054,China)

机构地区：[1]新疆农业职业技术学院,新疆昌吉831100 [2]新疆师范大学数学科学学院,新疆乌鲁木齐830054

出　　处：《数学的实践与认识》2023年第11期158-163,共6页Mathematics in Practice and Theory

基　　金：国家自然科学基金(11261061,61362039);新疆师范大学课题(XJNUSY082017B03);新疆农业职业技术学院课题(XJNZYKY201528)。

摘　　要：为解决C-V模型中弱边缘或边缘模糊图像分割问题,提出了用边缘停止函数代替正则化Dirac函数的C-V图像分割模型.首先对正则化Heaviside函数和正则化Dirac函数中的参数进行了讨论,然后利用图像边缘信息将梯度算子引入正则化Driac函数中,对C-V模型进行改进,最后,用边缘停止函数代替C-V模型中的正则化Dirac函数.实验结果显示,提出的模型比C-V模型对图像的分割效果更好.In order to solve the problem of image segmentation with weak edges or blurred edges in the C-V model,a C-V image segmentation model with edge stop function instead of regularized Dirac function is proposed.First,the parameters in the regularized Heaviside function and the regularized Dirac function are discussed,then,the gradient operator is introduced into the regularized Driac function to improve the C-V model,Finally,the regularized Dirac function in C-V model is replaced by edge stop function.Experimental results show that the proposed model is better than C-V model in image segmentation.

关 键 词：偏微分方程 图像分割 C-V模型 正则化Dirac函数 边缘停止函数 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]