期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2008
卷号：XXXVII Part B3a
页码：241-246
出版社：Copernicus Publications
摘要：Bayesian methods coupled with Markovian frameworks has several applications in remote sensing images processing, such as the pixel level applications like filtering, segmentation and classification, and the higher level applications like object recognition and organization etc. This article illustrates the powerfulness of Markovian model at two levels for the road extraction problem in remote sensing images. In order to obtain the final road network, one of the low level applications is using Gaussian Markovian model to segment road target in images and then treat with the original segmentation by the line segment match method and mathematical morphology. For the sake of renewing the complete road network, one of the high level applications detects the basic road sections by homogenous texture and line segment match method, and then organizes the basic road sections in combine with context through adopting Markovian model. Experimental results show that Markovian model has good road segmentation results and high ability to interpret road network
关键词：Markov Random Field; Bayesian Estimation; Road Segmentation
