ENGINEERING JOURNAL OF GEOSPATIAL INFORMATION TECHNOLOGY
Year:2017 | Volume:5 | Issue:1 |Papers Count:8

Cognition of travel mode and travel demand is of prime importance to transportation communities and agencies in every country. If the precise transportation modes of individual users are recognized, a more realistic travel demand can be considered. Also, in location-based service, the knowledge of a traveler’s transportation mode is applied to send targeted and customized informative advertisements. This study examines the feasibility of using a neuro-fuzzy inference system to automatically detect the mode of transportation from GPS data collected by GPS-enabled mobile phones. To achieve this, the knowledge was extracted in the form of fuzzy rules from the data and, then, the rules are being used for determination of transportation’s mode. For this purpose, the model was examined in two cases. In the first case, all GPS data from mobile devices were used, while in the second case the critical point algorithm was exercised. In addition to reducing the size of required GPS datasets, the critical point algorithm decreases data collection cost and saving mobile phone resources such as its battery life. The results showed that the suggested model have the capability of detecting a transportation mode with 94.1 percent accuracy in case of using all GPS data and 95.5 percent accuracy in case of using critical points.

In this study, a method for unsupervised change detection in multi-temporal SAR images has been presented based on integrating clustering and active contour model. In this method, texture information is extracted by using Gabor filter in different scales and directions. KPCA transformation is also applied to reduce the dependency between the extracted features and image information. Moreover, Discrete Wavelet Transformation (DWT) and Gustafson-Kessel clustering (GKC) methods are used respectively to generate the difference image and the initial contour for the active contour model. In the final step, the region-based nonparametric active contour model is used for producing the change image containing changed and unchanged regions. For performance evaluation of the proposed method, two sets of high resolution multi-temporal Terra SAR-X images are considered. Experimental results of unsupervised change detection method show that, the total error rate of the proposed approach for the first data set are decreased respectively to 4.95%, 3.30% and 3.34% compared to that of the Chan-Vese, MRF and EMMRF methods and for the second data set, the total error rate of the proposed method are decreased to 2.56%, 1.86% and 1.87 As well. Moreover, the results showed that the use of GKC method leads to production of the initial curve with minimal convergence time for the active contour model. Also, the use of active contour model improves the accuracy of change map creation using a repititive process.

Lack of interaction and ability to combine between ionosphere models are important problems among them.This study has investigated the interaction between global and regional ionospheric models which are according to the base mathematical functions. Here, spherical Slepian base functions have been suggested for spectral combination. They can be defined in global and regional scales. Two sets of coefficient were assumed for spherical Slepian base function. One set (direct model) is extracted from the direct observation of GPS stations. Another set (indirect model) is derived from spherical harmonic products of Global Ionosphere Models (GIMs (. GIM’s efficiency has been modified on a sub-space which was equal to pervious set.Maximum degree has considered equal to 15 for implementation of spherical Slepian spectral combination theory. Also, the Arctic region has been taken as study area where it is the spherical cap by latitude upper than 60 degrees. Observations of three GPS stations are used for evaluation of models. They have not contributed for direct and indirect modelling. Root Mean Square of Errors (RMSEs) for GIMs, direct, indirect and combining models are equal to 3.7, 2.2, 1.9, 1.4 TECU, respectively. In other words, results show that combining method has improved the ionospheric modelling here. The combining method can decrease the effects of the lack of stations in the indirect model and the inappropriate distribution of stations in the direct one. Therefore, the RMSE of the combining model is less than other models in the Arctic region.

Population growth, urbanization and immigration from rural areas to cities results in increasing expansion of the cities. The adequate urban development requires proper and accurate planning, such those facilities for the new areas that are provided and environmental impacts are lessened. Computer based simulation and prediction of urban growth can be assumed as a good start for urban planning. In this research, a new agent base simulation of urban growth is developed, in which the communication and decision of agents are imitated from the Particle Swarm Optimization (PSO) algorithm. The model is tested on the urban growth of Zanjan city- Iran between 2005 and 2015. In this model, land developers are classified into three groups of agents according to their income level. These agents search the environment and find proper lands for development according to their priorities and conditions. The output of the model is 74% similar to the reality according to the Kappa index. This and other results show that the model can predict the expansion of the city adequately.Moreover, the comparison made shows that modeling of the relations and communications between agents similar to PSO can slightly improve the quality of the model. The results showed the adequacy of the proposed agent-based modelling for the simulation of urban growth. To have a more accurate model, it is recommended to model the behavior of the agents with more details and also to consider the competition between agents.

Ocean tides cause noticeable aliasing errors in the gravity field from single pair space-borne graimetric missions like GRACE. Several studies about future gravity missions have shown that constellations with two or more GRACE-like tandems can lead to a significant reduction of aliasing error from all kinds of high frequency signal sources. Despite such reduction, tidal aliasing will remain an error source still. This study investigates the efficiency of tidal error dealiasing in the post-processing mode for such future dualpair missions. To that purpose, we analyze the way a certain satellite mission sampling each tidal constituent.Given the repeat orbit patterns and the observation time span, we examine and model the aliasing periods and amplitudes constitute by constitute. Results show that a double-pair formation has de-aliasing function comparing to a single-pair formation in terms of distribution and amplitude of ocean tide aliasing error. With least-square (LS) estimation, the aliasing error can be reduced significantly.

By increasing the spatial resolution of remote sensing imaging sensors, the image analyzing paradigm is moving towards the object based image analysis approaches, instead of single pixels. Among the common segmentation algorithms, super-pixel methods are presenting themselves as the new tools in computer vision. In this paper, the capabilities of a state-of-the-art super-pixel algorithm, namely called SLIC, is investigated for creating image segments from ultra-high resolution remote sensing images. In our proposed method, square and hexagonal super-pixels were formed and then DBSCAN clustering algorithm is employed to build image segments from these pixels. The results were compared to image segments obtained from FNEA algorithm, a well-known method for remote sensing image segmentation. Visual and quantitative evaluations demonstrate the efficiency of proposed method.

Quality and quantity of vegetation land cover is considered as one of the important aspects of environment.Detection of trends in natural phenomena such as vegetation, requires long-term studies, more than lifetime of a satellite. On the other hand, combining data from different sensors could lead to formation of false changes.One of the main causes of false changes is different spectral sensitivity functions (SRFs), among sensors under study. In this regard, the impact of these factors should be eliminated or reduced as much as possible by a procedure named relative calibration which is the main goal of this research. There are similarities between Landsat satellites series and SPOT-5 with Sentinel-2 in many aspects, so MSI (the Sentinel-2’s sensor) has capacity for data continuity. In this study, by incorporating polynomial equations, Landsat sensors (OLI, ETM +, ETM) and SPOT-5 were calibrated relative to MSI. The combination of radiative transfer models; PROSPECT-4 for leaf and 4SAIL for canopy, were used to simulate 50000 top of canopy synthetic spectral signatures and then soil effect was combined with them using linear spectral mixture model. After all, 150000 signatures were simulated. These spectral signatures were transformed to equivalent reflectance values (Blue, Red, NIR and SWIR) and spectral indices (NDVI, EVI and NDWI).80% of spectral signatures were selected randomly for solving relative calibration models. Also, for validation purpose, remained simulated (20%) and 38 top of canopy measured spectral signatures were used. According to the results, linear equation can model the difference (caused by SRF) between MSI and others quite well and there is no need for more complicated equations. In general, results of this research show high and acceptable correlation for all reflectance bands and indices. It is more necessary to perform a relative calibration pre-processing step for EVI time series. Amongst reflectance bands, NIR has the highest continuity.

Although, the distinction between the land cover classes was increased in large feature space of remote sensing images, but the low number of training data prevent this. In order to solve this problem, ensemble classification methods can be used instead of individual classifiers. In this paper, a new method for ensemble support vector machine was proposed called “Support Vector Random Machines (SVRMs)”. In proposed method, bootstrap was produced using modification of training data and feature space. Simultaneous boosting SVM was used for basic classifiers. Then, classification map was resulted using SVM fusion of basic classifier. Hyperspectral and Polarimetric SAR data was chosen for evaluation performance of the SVRMs. Experiments were evaluated from three different points of view: First, evaluation against other ensemble SVM methods; second, evaluation against various feature selection methods in classification and third, evaluation against the various basic and new classification methods. As the results, proposed method is 16% better than the individual SVM classifier in hyperspectral data and this is 10% in PolSAR data. Also, the classification results of SVRMs in various classes compared to other SVM ensemble method were improved. The results reported from the proposed method compared to the other feature selection method (Genetic Algorithm) has the effectual performance in classification.The results show that the proposed method presents a better performance compared to the basic classification methods (maximum likelihood and wishart) and advanced classification (random forest and neural network).