Extracting Roads from High-resolution RS Images Based on Shape Priors and Graph Cuts

Abstract

Abstract:

A new road extraction scheme is proposed. First of all, preprocess a RS image with a filter to obtain texture features; then calculate the classic energy function with Graph Cuts to obtain road points, and t-link is constructed with Gaussian Mixture Model; the initial road stripes are achieved with the Line Segment Match Method; lastly, calculate the energy function merged with two shape priors with Graph Cuts to connect roads, which combines the strategy of dynamic segmentation. Experiments of different images testify our algorithm’s effectiveness.

Key words: Road Extraction, Gaussian Mixture Model, Shape Prior, Graph Cuts

CLC Number:

P237

References

[1]Kolmogorov, V., Zabih, R.: What Energy Function Can Be Minimized via Graph Cuts? [J].IEEE Transaction on PAMI 26(2) (February 2004) 147–159.
[2]D. Cremers, S. Osher, S.Soatto. Kernel Density Estimation and Intrinsic Alignment for Shape Priors in Level Set Segmentation[J]. IJCV 69(3) (September 2006),335–351.
[3]Y.Chen, H.Tagare, S.Thiruvenkadam. Using Prior Shapes in Geometric Active Contours in a Variational Framework[J]. Int. J. of Computer Vision,50(3):315-328,2002.
[4]D.Cremers, N.Sochen, C.Schnorr. Towards Recognition-Based Variational Segmentation Using Shape Priors and Dynamic Labeling[C]. In L.Griffith, editor, Int. Conf. on Scale Space Theories in Computer Vision,volume 2695 of LNCS:388-400,Isle of Skye,2003,Springer.
[5]T.Chan, Wei Zhu. Level Set Based Shape Prior Segmentation[J]. Computer Vision and Pattern Recognition,2005,volume 2:2264-1170,2005.
[6]Mikael Rousson, Nikos Paragios. Shape Priors for Level Set Representations[J]. ECCV,2002,LNCS 2351,pp.78-92.
[7]M. Rousson and D. Cremers. Efficient Kernel Density Estimation of Shape and Intensity Priors for Level Set Segmentation[C].Proc.Int’l Conf. Medical Image Computing and Computer-Assisted Intervention, vol. 1, pp. 757-764, 2005.
[8]T.Schoenemann, D.Cremers: Globally Optimal Image Segmentation with an Elastic Shape Prior. In: ICCV. (2007) 1–6.
[9]D. Freedman, T. Zhang. Interactive Graph Cuts Based Segmentation with Shape Priors[J]. Proc of IEEE Computer Society Conference on Computer Vision and Pattern Recognition．Washington DC:IEEE Computer Society,2005:755-762．
[10]O. Veksler. Star Shape Prior for Graph-cut Image Segmentation[C]. Proc of the 10th European Conference on Computer Vision: Part III．Berlin: Springer-Verlag，2008:454-467．
[11]Slabaugh, G., Unal, G.: Graph Cuts Segmentation Using an Elliptical Shape Prior[J].In: ICIP. (2005) II: 1222–1225
[12]Das, P., Veksler, O., Zavadsky, S., Boykov, Y.: Semiautomatic Segmentation with Compact Shape Prior. In: CRV. (2006) 28–36.
[13]S.Vicente, V.Kolmogorov, C.Rother: Graph Cuts Based Image Segmentation with Connectivity Priors[J]. In: CVPR. (2008).
[14]Hui Wang, Hong Zhang. Adaptive Shape Prior in Graph Cuts Segmentation[J]. Image Processing,ICIP,3029-3032,2010.
[15]Nhat Vu, B.S.Manjunath. Shape Prior Segmentation of Multiple Objects with Graph Cuts[J]. Computer Vision and Pattern Recognition,2008:1-8.
[16]J. Malcolm, Y. Rathi, and A. Tannenbaum. Graph Cuts Segmentation with Nonlinear Shape Priors. In ICIP, 2007.
[17]L.R. Ford and D. R. Fulkerson, A Simple Algorithm for Finding Maximal Network Flows and an Application to the Hitchcock Problem, Canad. J. Math, 9(1957), 210-218.
[18]Yuri Boykov, Vladimir Kolmogorov. An Experimental Comparison of Min-Cut Max-Flow Algorithms for Energy Minimization in Vision[J].IEEE Transactions on PAMI.2004,26:1124-1137.
[19]Boykov, Y., Funka Lea, G.: Graph Cuts and E?cient N-D Image Segmentation[J].International Journal of Computer Vision 69(2) (September 2006) 109–131.
[20]LI JinHui.Segmentation Algorithm for Road-Images Using Gabor Filter and Markov Random Field[D]. HoHai University,2009.（黎瑾慧. 基于Gabor滤波器和MRF的道路影像分割方法研究[D].河海大学,2009.）
[21]SHI Wenzhong, ZHU Changing. The Line Segment Match Method for Extracting Road Network from High-Resolution Satellite Images[J].IEEE Transaction on Geoscience and Remote Sensing.2002,40(2):511-514.