Mahboobeh mohammad Yusefi Bohluli Ahmadi; Abdolreza Safari; َAnahita Shahbazi
Abstract
Abstract
Global gravity field is commonlymodelled in spherical harmonic basis functions to a certain degreeof spectral and spatial resolution. Non-uniformdistribution and different quality data limitthese functions in local gravity field modeling.Spherical harmonic basis functionsshow more global properties ...
Read More
Abstract
Global gravity field is commonlymodelled in spherical harmonic basis functions to a certain degreeof spectral and spatial resolution. Non-uniformdistribution and different quality data limitthese functions in local gravity field modeling.Spherical harmonic basis functionsshow more global properties that means they are suitable forshowing low frequency gravityfield. In local-scale studies, radial basis functionson the sphere with quasi-local support can improve gravityfields up to a high spatial/spectral resolution.The local modelsare usually moreaccurate than global modelsin the desired locations.These functions are usually notorthogonal on a sphere, which makes the modelling process morecomplex.In this study we evaluated the radial basis functions: point-mass kernel, radial multipoles, Poisson and Poisson wavelet ,and then we compared their performances in regional gravity fieldmodelling on the sphere using real gravity acceleration data in Farscoastal area. A least-squares technique has been used toestimate the gravity field parameters. Iterative Levenberg-Marquardtalgorithm is appliedfor nonlinear inverse problem solving and minimization of differences between calculated andobserved values. These parameters include number, location, depth and scalingcoefficients in radial basis function.In order to increase efficiency Levenberg-Marquardt algorithm for solving gravity field modeling, the initial valueof theregularization parameter determined with a relation based on objective functionJacobian and also a method is provided for this parameter updates. Theresults showed that the accuracy of gravity field modeling forany types of radial basis function would be almost thesame, if the depths of SRBFs are chosen properly.
Davood Akbari; Abdolreza Safari; Saeid Homayouni
Abstract
Abstract Recently, a new approach, based on the Hierarchical SEGmentation (HSEG), grown from automatically selected markers using Support Vector Machines (SVM), has been proposed for spectral-spatial classification of hyperspectral images. The HSEG algorithm, which combines region object finding with ...
Read More
Abstract Recently, a new approach, based on the Hierarchical SEGmentation (HSEG), grown from automatically selected markers using Support Vector Machines (SVM), has been proposed for spectral-spatial classification of hyperspectral images. The HSEG algorithm, which combines region object finding with region object clustering, has given good performances for hyperspectral image analysis. This technique produces at its output a hierarchical set of image segmentations. This paper aims at improving this approach by using image segmentation to integrate the spatial information into the marker selection process. In this study, the markers are extracted from the classification maps obtained by both SVM and watershed segmentation algorithm. The watershed algorithm is used in parallel and independently to segment the image. It is a powerful morphological approach to image segmentation. Moreover, the class’s pixels, with the largest population in the classification map, are kept for each region of the segmentation map. Lastly, the most reliable classified pixels are chosen from among the exiting pixels as markers. Then, a marker-based HSEG algorithm is applied. Each region from the segmentation map is classified by applying a majority vote rule over the pixel-wise SVM classification results. Three benchmark urban hyperspectral datasets are used for our comparisons: Pavia, Berlin and DC Mall. The results of our experiment indicate that, compared to the original hierarchical approach, the marker selection using segmentation algorithm leads in more accurate classification maps. Indeed, the proposed approach achieves an approximately 4%, 6% and 5% kappa coefficient higher than the original hierarchical-based algorithm for the Pavia, Berlin, and DC Mall datasets, respectively.
Fereydoon Nobakht Ersi; Abdolreza Safari; Mohammad Ali Sharifi
Abstract
The main purpose of the present paper is to use the ARMA probability models to model the time series of the daily positions of GPS permanent station.Daily Locations of the LLAS permanent station in the Southern California region have been selected from the SCIGN network, covering a period of seven years ...
Read More
The main purpose of the present paper is to use the ARMA probability models to model the time series of the daily positions of GPS permanent station.Daily Locations of the LLAS permanent station in the Southern California region have been selected from the SCIGN network, covering a period of seven years from January 2000 to December 2006, to establish a time series of position and to analyze it. Based on the time series of the daily position and using the weighted least squares, the geodetic parameters such as linear trend, annual and semi-annualfluctuations, as well as offsets,have been simultaneously estimated for the LLAS permanent station. In this study, Auto correlation Functions (ACF) and Partial Auto Correction functions (PACF) are used as the study tools for identifying the time series behavior of daily position of GPS permanent station and provide the possibility to examine the dependency of the position time series daily data. Given that several different probabilistic models may be appropriate for a daily position time series, therefore,the Akaike Information Criterion has been used at the stage of identifying and selecting the useful model. In this study, numerical results show that the best autoregressive moving average (ARMA) probabilistic model for the LLAS permanent station is ARMA (1, 1) for direction N. Also, the ARMA (2, 1) probabilistic model is the most appropriate model for direction E, while the ARMA (1, 2) probabilistic model is the best model for direction U. After estimating an appropriate probabilistic model for the time series of the daily position of the GPS permanent station, it is possible to predict the time series of the position along with the trend and seasonal components.
