基于引导滤波的多阈值图像分割算法研究  

Research on multi-threshold image segmentation algorithm based on guided filtering

作　　者：苗田田 李昊宸 MIAO Tiantian;LI Haochen(School of Infomation Engineering Zhengzhou University of Science and Technology,Zhengzhou,Henan,450064,China)

机构地区：[1]郑州科技学院信息工程学院,河南郑州450064

出　　处：《长江信息通信》2024年第4期58-60,共3页Changjiang Information & Communications

摘　　要：由于缺少对图像背景信息的识别与剔除,导致图像分割精度较低,分割效果不佳。为此,提出基于引导滤波的多阈值图像分割算法研究。采用引导滤波算法对图像像素点进行平滑处理,根据多线性图像特征属性对图像去噪,对图像的背景信息进行识别与剔除,提取图像区域级信息,以此为依据,修正处理图像隶属度矩阵,得到新的隶属度矩阵与最佳分割阈值,实现图像多阈值分割。实验结果表明,所研究算法对于图像分割具有较高的精度,分割效果较好。Due to the lack of recognition and elimination of image background information,the image segmentation accuracy is low and the segmentation effect is not good.Therefore,a multithreshold image segmentation algorithm based on guided filtering is proposed.Thc guided filtering algorithm is used to smooth the image pixels,denoise the image according to the feature attributes of multi-linear images,identify and climinate the background information of the image,and extract the regional information of the image.Based on this,the membership matrix of the image is modified to obtain a new membership matrix and the optimal segmentation threshold,and the multi-threshold segmentation of the image is realized.The experimental results show that the algorithm has high accuracy and good segmentation effect.

关 键 词：引导滤波 多阈值 图像分割 修正处理 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]