北京测绘2025，Vol.39Issue(3)：370-375,6.DOI:10.19580/j.cnki.1007-3000.2025.03.021
基于MRF模型的高层建筑立面测绘影像重复纹理分割方法
A repetitive texture segmentation method for facade surveying and mapping images of high-rise buildings based on MRF model
摘要
Abstract
In order to improve the accuracy of repetitive texture segmentation in facade surveying and mapping images of high-rise buildings and enhance image quality,a repetitive texture segmentation method for facade surveying and mapping images of high-rise buildings based on the Markov random field(MRF)model was studied.Fourier transform was used for phase matching of facade surveying and mapping images of high-rise buildings.The repetitive texture was transformed and processed in the frequency domain,and the translation difference between images was calculated.Based on the Fourier transform and image grayscale correlation coefficient,repetitive textures in the horizontal and vertical directions of high-rise building images were detected,and regional and detailed information of repetitive textures in the images was obtained.A high-order MRF model was developed,and the region information,detail information,and prior energy of repetitive textures in the images were integrated.The optimal marking field for repetitive textures in the images was derived to complete the repetitive texture segmentation in facade surveying and mapping images of high-rise buildings.The experimental results show that the method has strong anti-interference performance in detecting repetitive textures in facade surveying and mapping images of high-rise buildings and high segmentation accuracy in repetitive textures in the images,and it can retain the detailed information of segmented images,improving image quality.关键词
马尔可夫随机场(MRF)模型/高层建筑/立面测绘影像/重复纹理/分割方法Key words
Markow random field(MRF)model/high-rise building/facade surveying and mapping image/repetitive texture/segmentation method分类
天文与地球科学引用本文复制引用
王云富..基于MRF模型的高层建筑立面测绘影像重复纹理分割方法[J].北京测绘,2025,39(3):370-375,6.基金项目
江苏省高等教育教改立项研究课题(2017JSJG485) （2017JSJG485）
