针对传统的超像素马尔科夫随机场(Markov random field,MRF)影像分割模型中对空间背景信息利用不够完全的问题,发展了一种新的超像素MRF模型。该算法将高阶邻域模型引入到MRF的交互势函数中,使交互势函数能够充分利用超像素邻域系统所包含的空间背景信息。结合此一阶势函数模型,还提出一种逐类别的β参数自动估计方法,该方法是在范数距离的基础上进行的。利用2景具有不同特点的农田地区高分遥感影像,开展了验证实验。实验结果表明,本算法对于边界强度等空间背景信息的利用效果更好,分割结果更精确。与其他超像素MRF分割算法对比,也说明了该算法在性能上的优越性。
In view of the problem that the traditional super-pixel Markov random field (MRF) image segmentation model cannot fully utilize spatial context information, a new super-pixel MRF model is proposed. This algorithm incorporates higher-order neighborhood model into the interactive potential term of MRF. The new model enables the interactive potential to fully exploit the spatial context information contained in the super-pixel neighborhood system. Additionally, a new class-wise estimation method for β is proposed, which is based on norm distance. By utilizing two scenes of high-resolution remote sensing images acquired over different agricultural landscapes, validation experiment was conducted. The experiment results indicate that the proposed method can better use the contextual information such as edge strength, thus achieving higher segmentation accuracy. Moreover, the algorithm proposed by the authors showed superior performance when it was compared with other super-pixel MRF approaches.
Löw F,Conrad C,Michel U.Decision fusion and non-parametric classifiers for land use mapping using multi-temporal RapidEye data[J].ISPRS Journal of Photogrammetry and Remote Sensing,2015,108:191-204.
[2]
Kim H O,Yeom J M.Effect of red-edge and texture features for object-based paddy rice crop classification using RapidEye multi-spectral satellite image data[J].International Journal of Remote Sensing,2014,35(19):7046-7068.
Yang Y J,Zhan Y L,Tian Q J,et al.Crop classification based on GF-1/WFV NDVI time series[J].Transactions of the Chinese Society of Agricultural Engineering,2015,31(24):155-161.
[4]
Liu M W,Ozdogan M,Zhu X J.Crop type classification by simultaneous use of satellite images of different resolutions[J].IEEE Transactions on Geoscience and Remote Sensing,2014,52(6):3637-3649.
[5]
Besag J.On the statistical analysis of dirty pictures[J].Journal of the Royal Statistical Society,Series B Methodological,1986,48(3):259-302.
[6]
Kumar S,Hebert M.Discriminative random fields[J].International Journal of Computer Vision,2006,68(2):179-201.
Liu L,Shi Z G,Su H R,et al.Image segmentation based on higher order Markov random field[J].Journal of Computer Research and Development,2013,50(9):1933-1942.
[9]
Yu Q Y,Clausi D A.IRGS:Image segmentation using edge penalties and region growing[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2008,30(12):2126-2139.
[10]
Qin A K,Clausi D A.Multivariate image segmentation using semantic region growing with adaptive edge penalty[J].IEEE Transactions on Image Processing,2010,19(8):2157-2170.
Su T F,Li H Y.GBM-combined MRF method for remote sensing image segmentation[J].Journal of Inner Mongolia Agricultural University(Natural Science Edition),2015,36(1):143-149.
[12]
Achanta R,Shaji A,Smith K,et al.SLIC superpixels compared to state-of-the-art superpixel methods[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2012,34(11):2274-2282.
[13]
Comaniciu D,Meer P.Mean shift:A robust approach toward feature space analysis[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2002,24(5):603-619.
[14]
Levinshtein A,Stere A,Kutulakos K N,et al.TurboPixels:Fast superpixels using geometric flows[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2009,31(12):2290-2297.
[15]
Su T F,Li H Y,Zhang S W,et al.Image segmentation using mean shift for extracting croplands from high-resolution remote sensing imagery[J].Remote Sensing Letters,2015,6(12):952-961.