ENGINEERING JOURNAL OF GEOSPATIAL INFORMATION TECHNOLOGY
Year:2018 | Volume:6 | Issue:2 |Papers Count:10

Classification is the most common method for information extraction from remotely sensed images. Conventional classification methods are mostly based on spectral information. While particularly in high spatial resolution images, spatial relationships between neighboring pixels used to discriminate between different land-cover classes in human interpretation. In different research methods for quantification of image texture and use it to create separation between classes is provided. Considering the variety of formulation and adjustable parameters of texture quantization methods, huge number of texture features could be generated. Each feature has specific ability to discriminate special classes. In this paper a new method based on spectral index formulation proposed to generate Spatial Indices from textural features. Best pairs of textural features selected and Spatial Indices using them will be generated. Generated Spatial Indices are good abstract of textural feature space to use in classification procedure. This method could led to better classification results in a direct and none international solution.

Monitoring Land use changes is one of the important applications of remote sensing and geographic information system. In this study, a framework for change monitoring in multitemporal satellite images is presented by Iteratively Reweighted multivariate alteration detection (IR-MAD) algorithm and support vector machine (SVM) classification. In this study, the change detection analysis has been done using multitemporal Landsat satellite images with 18 years time interval of Shahi Island and a part of the western region of Lake Urmia. The proposed method has two main steps in change monitoring. In the first step, components of change intensities are determined automatically by IR-MAD transformation. In the following, optimized components are selected by applying the kernel principal component analysis (KPCA) on components of change intensities. In the next step, for generating the content of change map, The combination of optimal components is classified by SVM method. For the evaluation performance of the proposed method, in change monitoring, this method was compared with conventional methods such as analysis of the spectral– temporal combination and post classification comparison. The experimental results show that the overall accuracy of the proposed method increased 4. 89% and 4. 39% compared to that of the spectral-temporal Combination and post classification comparison, respectively.

Geographically Weighted Logistic Regression (GWLR) is a local version of logistic regression (LR) which estimates different relationships between independent and dependent variables at each location. In this research, local model (GWLR) is used for defining CA transition rules and evaluating GWLR capabilities in terms of enhancing urban development prediction accuracy. Also, a new parameter named “ Edge Expansion Coefficient” was defined for the determination of tradeoff between two important urban development processes: edge expansion and spontaneous growth. Moreover, in order to assess the prediction accuracy, fuzzy Kappa statistic was applied along with the traditional Kappa coefficiency. The developed CA model in this study was run for the prediction of urban development in south west of Tehran metropolitan area during 2004-2013 period. The results of the study showed that, using GWLR model for defining CA‟ s transition rules, one can significantly increase urban development prediction‟ s accuracy compared to that of predicted urban development by CA model based on logistic regression (Logistic-CA). The prediction accuracies of the proposed model in this research and the Logistic-CA were 0. 54 and 0. 30, respectively, as measured by Kappa coefficient. Also, the prediction accuracies of the proposed model were calculated to be 0. 68 and 0. 76 when measured in terms of fuzzy Kappa statistic with halving distances of 50 and 100 meters in exponential distance decay function, respectively.

Three dimensional models generated from cities, due to their wide range of applications, are required by most organizations and institutions. Among numerous 3D reconstruction techniques of urban areas, stereo-panorama images are largely taken into consideration for their full coverage of the environment. In this paper, a new stereo panorama scanner is presented which utilizes stereo cameras. The system provides the possibility to extract accurate geometric information for 3D reconstruction of urban areas and stereo viewing of the surroundings as well. The advantages of such system are: the automation of matching and 3D modelling process, equal resolution over the whole stereo panorama and the ability to extract precise 3D models in pointcloud format. In this paper, the structure of the system is described, including system implementation, calibration and 3D model generation. Additionally, experimental results of system evaluation are provided. The evaluation results show that the system is enable to generate accurate models for 3D city reconstruction applications. The results demonstrate that once the distance from stereo scanner to the object is in the range of some 15 meters, consequently the discrepancy of the point distances would be around 57 millimetres.

One of the main challenges of human is the dramatic decrease in resources due to human‟ s excessive consumption of land that has led to a phenomenon called land degradation. Various models have thus far been introduced for assessment of this phenomenon. The parameters and their weights differ from one model to another as per experts‟ opinion. The present study introduces a new method to identify and calibrate the parameters, as per the conditions of the region under study, affecting this phenomenon. The proposed method is considered as a data-based model such that parameter weights are computed intelligently and as per the climate and geographical conditions of the region. The genetic algorithm and Weighted Overlay Index were used to determine the significance level and ranking of the criteria. For the purpose of assessment, the data pertaining to Neinava region, located in Iraq, including Landsat satellite images of 1985, 2001, and 2014 as well as criteria such as distance from rivers, distance from lakes, distance from agricultural areas, distance from roads, distance from residential areas, height, slope, distance from Qanats, distance from wells, erosion, type of climate, and NDVI index were used. The results obtained from modeling and calibration as per the proposed model were compared with those of the regular method (application of equal weights). Application of genetic algorithm and calibration of weights yielded a standard deviation of 0. 03 for prediction of vegetation degradation which is considerably lower than that yielded by the regular method (0. 137). The criteria were also prioritized at this stage as per their significance. To ensure the model accuracy, data of 2001 and 2014 were used to assess the obtained results. The assessment result yielded a standard deviation of 0. 053 and accuracy of 0. 857. After the accuracy of the model was ensured, the vegetation degradation was predicted for 2027. The average rate of decreased NDVI values indicates the critical status of land degradation in the region under study.

Identification of objects referring to the same entity in different datasets is known as objects matching, which is both directly and indirectly used in a wide range of applications including conflation, quality assessment, data updating, and multi-scale analysis. Hence, a novel object matching approach is presented in this article, in which, in addition to take only geometric property into account, i. e. geometric and topological criteria, extracted from objects, any initial dependency on empirical parameters such as threshold of spatial similarity degree, buffer distance, and metric weights is eliminated, through which matching procedure may then be conducted in different datasets. All the relations in the proposed approach are considered including: one-tonull, null-to-one, one-to-one, one-to-many, many-to-one, and many-to-many. Moreover, efficiency of linear object matching using Real Coded Genetic Algorithm (RCGA), Particle Swarm Optimization (PSO) algorithm, and Artificial Bee Colony (ABC) algorithm in different datasets were investigated through optimization of geometric criteria. In order to assess the efficiency of the proposed approach, three datasets of different scales from various sources were used. As indicated by the results, the proposed framework was able to appropriately identify corresponding objects in different datasets. Additionally, it was revealed that GA outperformed the other two algorithms in terms of optimizing the parameters present in linear object matching.

Today, Industry and traffic can be a major contributor to the air pollution in the cities. The traditional methods for the study of air pollution are based on chemical measurements and analysis which requires time, labor along with relatively high costs. Study of spectral behavior of Plants affected by environmental stresses is one of the non-destructive methods in remote sensing science. The visible and near-infrared spectroscopy of plants technique, since it‟ s quick, easy to use and precise, is widely used to predict the biochemical components of plants and their changes. The aim of this study is to study the spectral reflectance behavior of leaves exposed to traffic pollution of a part of Imam Khomeini highway, Isfahan-Iran. Spectral characteristics of the leaf surface of infected species including ash, cypress and elm using spectral indices sensitive to stress and chlorophyll were studied. The results of artificial neural network to distinguish the control and polluted species using spectral indices (PRI, NDVI, Gitelson and … ( shows the accuracy of 73. 4%. The PLS regression model was conducted simultaneously for three species in polluted and control modes and it does not have any acceptable results. In addition, the model was conducted separately for each species; it showed acceptable results for ash trees in polluted and control modes.

Monitoring and surveillance changes around the world need powerful methods and techniques; consequently, detection, visualization, and assessment of significant changes are essential for planning and management. Incorporating PolSAR images as the result of interactions between electromagnetic waves and target due to a high spatial resolution almost one meter can be incorporated for studying changes on the Earth's surface. Analyzing full-polarized radar images comparing to single polarized radar images used amplitude and phase information of the surface in different available polarization (HH, HV, and VV). This study is based on the decomposition of full-polarized airborne UAVSAR images and integration of these features with algebra, similarity, and distance based methods for change detecting purposes using two real datasets. Assessing the accuracy of the method is implemented using ground truth data and different criteria for evaluating such as overall accuracy (OA), area under ROC curve (AUC) and false alarms rate (FAR). The output results showed that algebra change detection method has superiority to detect changes comparing to other implemented methods. Also, numerical results showed the superiority of algebra change detection algorithms comparing to others.

Biophysical characteristics and surface temperature are the key parameters to monitor and evaluate the physical and chemical processes of the earth's surface. The aim of this study is to investigate the relationship between urban heat island intensity and different biophysical characteristics in built-up and non-built-up lands. For this purpose, the four Landsat 8 satellite images and MODIS water vapor product in dates of April 10th 2016, June 29th 2016, August 27th 2014 and October 29th 2016 for cities of Ghaemshahr, Sari, Neka and Behshahr have been used. Single-channel algorithm was used to calculate the land surface temperature. A band combination tasseled cap extraction was used to estimate some surface biophysical characteristics as well. Also, a new method has been used to extract built-up regions. Finally, LST and biophysical characteristics for built-up and non-built-up regions were analyzed and then, correlations between them and heat island intensity have been surveyed in four months. The result showed that correlation coefficient of relationships between LST and biophysical characteristics surface was 0/88. Increase of differences between biophysical properties in built-up and non-built-up regions causes increase of the temperature difference between the two types of regions and so, intensification of urban heat island. For this reason, temperature difference between built-up and non-built-up regions varied between 0. 3-st and lowest heat island intensity corresponding to Ghaemshahr and Neka were estimated 0. 5768 and 0. 03 respectively.

Topographic masses above the geoid are considered as a major obstacle in geoid determination by using Global Gravitational Models (GGMs). GGMs provide the possibility of the Earth's potential field modeling as the expansion of the external-type series of spherical harmonics. Applying the external expansion to obtain disturbing potential on the geoid within the topographic masses will cause a bias called „ topographic bias‟ . This study deals with calculating geoidal height using Earth Gravitational Model 2008 (EGM08). In order to do so, two methods of Direct Analytical Continuation one and Rapp's Indirect one are utilized. The Analytical Continuation Approach is based on using EGM08 within the topographic masses and applying topographic bias. Alternatively, Rapp‟ s Approach is based on calculating height anomaly and its downward continuation on the geoid. The success of these two methods to geoid simulation on 490 GPS-Levelling stations in mountainous region of Colorado in the USA were evaluated. The results are an indicator of the fact that two methods are compatible with each other with centimetric accuracy compared to GPS-Levelling points. Also, it suggests an improvement in the relative and absolute accuracy of the geoidal height resulting from EGM08 about 60% in both methods. The numerical investigation revealed that taking advantage of height harmonic models instead of point actual height can bring a bias in the matter of a few centimeters on the geoid. Moreover, the absolute accuracy of Rapp's Approach is higher than Analytical Continuation Approach in geoid determination in comparison GPS-Levelling points.