DEM fusion concept based on the LS-SVM cokriging method

ABSTRACT Data fusion from two sources of data could develop better output since the process may minimise any inherent disadvantages of the data. Cokriging data fusion requires a semivariogram fitting process, which is an important step for weight determination in the fusion process. The traditional method of cokriging fusion usually applies a specific model of semivariogram fitting based on the available options, such as circular or tetraspherical. This research aims to fuse height point data from two different sources using ordinary kriging based on LS-SVM regression, which is applied to the semivariogram fitting process. The data used are height points generated from stereo satellite imagery, GPS measurement, and topographic map points to generate DEMs. The research experiment begins by calculating the semivariogram model for all the data, and then the fitting process is performed by applying the same approach of functional approach for both sets of data. The following process is an ordinary cokriging interpolation, whose results are analysed and compared to the ordinary kriging interpolation. The experiment results prove that the ordinary cokriging fusion process could reduce interpolation error. The LS-SVM approach offers better precision in the semivariogram modelling by determining more precise weight calculation for cokriging fusion process.

[1]  Piazza Leonardo da Vinci,et al.  OPTICAL AND RADAR DATA FUSION FOR DEM GENERATION , 1998 .

[2]  KONRAD SCHINDLER,et al.  Improving Wide-Area DEMs Through Data Fusion – Chances and Limits , 2011 .

[3]  Choen Kim,et al.  A study on improving DEM-based forest biomass estimation using PALSAR and forest inventory data , 2014, 2014 IEEE Geoscience and Remote Sensing Symposium.

[4]  Stephen R. Marsland,et al.  Machine Learning - An Algorithmic Perspective , 2009, Chapman and Hall / CRC machine learning and pattern recognition series.

[5]  P. V. Arun.  A comparative analysis of different DEM interpolation methods , 2013 .

[6]  Coal Seam Thickness Prediction based on Least Squares Support Vector Machines and Kriging Method , 2015 .

[7]  P. Burrough,et al.  Principles of geographical information systems , 1998 .

[8]  Radford M. Neal.  Pattern Recognition and Machine Learning , 2007, Technometrics.

[9]  Phaedon C. Kyriakidis,et al.  Fusion of MODIS Images Using Kriging With External Drift , 2013, IEEE Transactions on Geoscience and Remote Sensing.

[10]  Jörgen Wallerman,et al.  Estimating biomass and height using DSM from satellite data and DEM from high-resolution laser scanning data , 2012, 2012 IEEE International Geoscience and Remote Sensing Symposium.

[11]  Helena Mitasova,et al.  Fusion of high-resolution DEMs for water flow modeling , 2017, Open Geospatial Data, Software and Standards.

[12]  Cheng Jiang,et al.  Fusion of photogrammetric and photoclinometric information for high-resolution DEMs from Mars in-orbit imagery , 2017 .

[13]  Konrad Schindler,et al.  Fusion of Digital Elevation Models Using Sparse Representations , 2011, PIA.

[14]  David M. Mount,et al.  Image Fusion Using Cokriging , 2006, 2006 IEEE International Symposium on Geoscience and Remote Sensing.

[15]  E. Baltsavias,et al.  A FRAMEWORK FOR THE FUSION OF DIGITAL ELEVATION MODELS , 2008 .

[16]  Qingmin Meng,et al.  High-resolution satellite image fusion using regression kriging , 2010 .

[17]  Shashi Kumar,et al.  PsInSAR based land subsidence estimation of Burgan oil field using TerraSAR-X data , 2018 .

[18]  A. M. Arymurthy,et al.  Accuracy Analysis of DEM Generated from Cokriging Interpolators , 2017 .

[19]  Daria Martinoni,et al.  A Distributed Geoprocessing Concept for Enhancing Terrain Analysis for Environmental Modeling , 2001, Trans. GIS.

[20]  Xiao Xiang Zhu,et al.  Fusion of TanDEM-X and Cartosat-1 DEMS using TV-norm regularization and ANN-predicted weights , 2017, 2017 IEEE International Geoscience and Remote Sensing Symposium (IGARSS).

[21]  Zhisong Huang,et al.  An Improved Kriging Interpolation Technique Based on SVM and Its Recovery Experiment in Oceanic Missing Data , 2012 .

[22]  Arthur L. Samuel,et al.  Some Studies in Machine Learning Using the Game of Checkers , 1967, IBM J. Res. Dev..

[24]  Aniati Murni Arymurthy,et al.  Semivariogram fitting based on SVM and GPR for DEM interpolation , 2019 .

[25]  N. Memarsadeghi,et al.  Efficient Kriging via Fast Matrix-Vector Products , 2008, 2008 IEEE Aerospace Conference.

[26]  Sang Hoon Lee,et al.  Multi-temporal, multi-resolution data fusion for Antarctica DEM determination using InSAR and altimetry , 2005, Proceedings. 2005 IEEE International Geoscience and Remote Sensing Symposium, 2005. IGARSS '05..

[27]  Johan A. K. Suykens,et al.  Optimal control by least squares support vector machines , 2001, Neural Networks.

[28]  Dodi Sudiana,et al.  Quality Assessment of DEM Generated from SAR Radargrammetry Based on Cross-Correlation and Spatial Resolution Setting , 2019 .

[29]  Kui Zhang,et al.  Mapping land subsidence in Jakarta, Indonesia using persistent scatterer interferometry (PSI) technique with ALOS PALSAR , 2012, Int. J. Appl. Earth Obs. Geoinformation.

[30]  Thierry Toutin,et al.  Stereo Radargrammetry With Radarsat-2 in the Canadian Arctic , 2013, IEEE Transactions on Geoscience and Remote Sensing.

[31]  Alexei Pozdnoukhov,et al.  Machine Learning for Spatial Environmental Data: Theory, Applications, and Software , 2009 .

[32]  Wenjiang Huang,et al.  Effects of different sampling densities on geographically weighted regression kriging for predicting soil organic carbon , 2017 .