ENGINEERING JOURNAL OF GEOSPATIAL INFORMATION TECHNOLOGY
Year:2019 | Volume:7 | Issue:2 |Papers Count:12

Nowadays, due to quality development of satellite images, automatic target detection on these images has been attracted many researchers' attention. Remote-sensing images follow various geospatial targets; these targets are generally man-made and have a distinctive structure from their surrounding areas. Different methods have been developed for automatic target detection. In most of these methods, target searching was used for features extraction or matching in order to detect the geospatial targets. Hence, in this paper, in order to improve computational time in target detection process, the areas of images with high probability of geospatial target existence, were selected. This will significantly improve the automation level of the process and computational time in following processes. For this purpose, a combination of saliency models and the of-local features detector's algorithm were used. The proposed method consists of three main steps, including local feature extraction by applying MSER algorithm, saliency maps generation by applying AWS and WMAP models and regions determination with high probability of geospatial target existence. In this paper, a threshold was defined by calculating the saliency values in the whole image. Salient regions were detected by applying the threshold for each extracted area. This method was implemented on six satellite images by different sensors and the mentioned six satellite images derived from Google Earth software that consist several geospatial targets with various backgrounds. Two criteria were used in order to quantitatively assess the results. Moreover, this method was compared with the mean shift segmentation algorithm in the color space images that was applied to detect saliency regions of an image. The results of the proposed method showed that the average of the detected areas was 5. 1% of the total area of the images in which 98. 28% of the targets were located in these regions. In addition, the average amount of computational time was 22. 1 seconds. The results showed the superiority of the proposed method in terms of accuracy and computational speed.

The effect of topography on the radiance record in satellite image, probably reduce the accuracy of algorithem impliementation on the images. Therefore, to reduce the effect of topography, various correction models based on interaction between light and object needs to be defined. This research introduces lombertin correction model (Cosine model) and non_lombertin correction model (mineart and C correction models) which can be applied on the serios landsat images of Irankouh. In order to evaluate of models, statistical parameters (mean difference and standard deviation of each band) before and after correction, Spectroscopy parameters (similarity angle and Euclidean distance) and accuracy of lead content estimation from Images were used. According to the results of the methods evaluation, the cosines model showed a poor performance compared to the other models. Overall accuracy of similarity angle (above 0. 97) between image and spectroscopy data for Minaret and Cosines models was 0. 973 and 0. 891, respectively. Also RMSE=0. 83 and R2 =0. 65 of lead content estimation of landsat 8 correction image using mineart model showed good performance.

Drilling requires accurate information about locations of underground infrastructures or it can cause serious damages. Augmented Reality (AR) as a technology in Ubiquitous GIS (UBIGIS) can be used to visualize underground infrastructures on smartphones. Since smartphone’ s sensors do not provide such accuracy, another approaches should be applied. Vision based computer vision systems are well known approaches to collect data of camera poses. Some vision systems track objects by image processing of natural environments which is quite difficult in unknown environments. Most vision systems are designed to detect targets. This paper aims to compare different types of targets that can be used in AR applications in UBGIS. For this propose, first three types of targets are being overviewed, then targets are compared in terms of their square or circular shapes. Finally, two kinds of targets related to Agisoft and Australis modeling software are chosen and images are being taken in different sincidence angles related to center of images at the area of faculty of Geodesy and Geomatics Engineering Khaje Nasir Toosi University of Technology. Then automatic recognition results in both software tools were being compared and analyzed. Comparisons reveal that QRCode is not appropriate due to its square shape and its distance limitation to scan by reader, because in AR applications, targets should be placed in a large field of view. Also, squared targets are not very suitable because their recognition is not easy. Therefore, Fiducial Markers are not recommended. It is worth mentioning that circular targets are the best target for this application. Among circular targets, although detection of circular Coded Target is not easy especially in very low incidence angle, but unrelated objects are not recognized as Coded Target; while detection of simple circular Target will make lots of mistakes. Generally, choosing appropriate target depends on corresponding application and usage. Then sensor-based and vision-based approaches to visualize underground infrastructures are compared. The results demonstrate that vision-based approach improve the precision of pose estimation parameters 8. 10260 m in position parameters and 10. 36 degrees in orientation parameters. Therefore, targets can be used as a vision-based approach.

Attainment of absolute deformation using Synthetic aperture radar interferometry (InSAR) requires the introduction of a reference point or area. In this study, geodynamic stations have been used to introduce this point or area and determine the absolute deformation. But considering that none of the permanent scatterers (PS) matches the geodynamic stations, there is no possibility of introducing a particular PS as the reference point. Hence, various geometric designs of scatterers were considered as reference and the influence of these designs on absolute deformation estimations was investigated. These designs include selecting the nearest PS to each geodynamic station, the average of permanent scatterers at a specific distance from the stations, the weighted average of all the permanent scatterers in the region and interpolation of deformation in the position of the stations. Furthermore, in order to investigate the effect of physical conditions and location of the reference points on absolute deformation, these designs were implemented on each of the geodynamic stations. The results indicate different absolute deformations relative to different reference points and using different geometric designs. The maximum root mean square error (RMSE) of the difference between InSAR and GNSS results of the geodynamic stations absolute velocity field along the line of sight is obtained by selecting the nearest PS to GTCL station as the reference point, which is about 3. 5 millimeters. Also the least RMSE is obtained by selecting weighted average of all permanent scatterers as reference which is less than 1 millimeter for all of the geodynamic stations.

Normalization of the surface temperature relative to environmental parameters is essential in scientific studies and urban and non-urban areas management. The aim of the current study is to propose a variance-based model for normalization of the surface temperature relative to environmental parameters. For this aim, Landsat 8 satellite bands, MODIS water vapor product, and ASTER digital elevation model were used. In this study, topography parameters, downward radiation on the surface, albedo, environmental lapse rate, vegetation and biophysical characteristics of the surface were considered as environmental parameters. Single channel algorithm was used for surface temperature calculation, and also an improved Coolbaugh model was suggested for modeling the downward solar radiation. Additionally, for modeling of albedo, environmental lapse rate and biophysical characteristics, a combination of Landsat 8 reflective bands, the digital elevation model, and tasseled cap transformation were exploited, respectively. Finally, the least square method was used to calculate the unknown coefficients of each parameter in the proposed normalized model, in order to minimiz the variance of the land surface temperature image. Coefficient correlation indexes and RMSE were used for accuracy assessment between the modeled and observed surface temperature values. Also the variance of normalized surface temperature image was used to estimate the proposed model capability. The results indicate that downward radiation on the surface parameter and both the elevation and greenness parameters, had the highest and the lowest effects on the surface temperature variation. The coefficient correlation and RMSE between the modeled and observed surface temperatures are 0. 97 and 1. 53, respectively and the variance of normalized surface temperature values is equal to 0. 79. Results of the current study implied the high efficiency of the proposed model for normalizing the land surface temperature relative to environmental parameters.

Surface albedo is one of the most effective practical remote sensing data being used directly or indirectly in different studies. In this study, various methods of atmospheric correction were investigated to pick up an accurate and precise method for estimating the surface albedo with atmospheric correction approach. Needless to say that information of humidity and digital elevation model are fundamental inputs in selected method which are being used automatically for every studying area. Then after running the model, results evaluated in two steps. At the first step, corrected atmospheric reflectance of OLI images was compared with MODIS BRDF product (MACD43A4). The results had a high correlation (between 0. 95 and 0. 97), and a smaller error (about 5. 9%) and is reduced absolute error rate in all bands about 24. 6%. In the second step, the broad band albedo was evaluated. Initially, the broadband albedo was compared to MODIS product (MCD43A3) in white and black sky conditions, with the lowest mean square error 0. 049 (WSA) and 0. 666 (BSA), successively. Moreover, this factor (broad band albedo) was assessed in different land uses, in which the maximum rate of broad band albedo changes before and after atmosphere correction relates to bright soil, dark mountainous and agricultural areas, respectively. The rate of albedo change in all land uses were decreased approximately between 0. 05 and 0. 18. Finally, it is suggested that, atmospheric correction is necessary in the surface-albedo estimation, mainly, it is essential in both automatic run and large areas with different climate characteristics.

Atmospheric correction of satellite images is important when vegetation indices are used to monitor changes. In this study, four methods of atmospheric correction were evaluated and validated using vegetation indices for monitoring vegetation. For this purpose, vegetation cover was measured at 19 points at intervals of 400-1000 m along a 10 km transect with 5 quadrats per point (95 quadrats in each period and 380 quadrats in total). Then, the synchronous images to the sampling dates in four correction methods including 1) QUick Atmospheric Correction (QUAC) 2) Fast Line-of-sight Atmospheric Analysis of Hypercubes (FLAASH) 3) Image Normalization of Iteratively Reweighted Multivariate Alteration Detection (IR-MAD) and 4) converting of digital numbers to Top-Of-Atmospher (TOA) reflectance techniques were applied. After that, Normalized Diffrences Vegetation Indx (NDVI) and Atmospheric Resistant Vegetation Indx (ARVI) were calculated. Next, the validation of linear regression models for the relationship between vegetation cover and vegetation indices with the two aforementioned vegetation indices was carried out based on 33 percent of testing data. Correlation coefficient (R) and R-squared (R2), Root Mean Square Error (RMSE), Absolute Mean Error (AME) and Bias were calculated as validity measures for each method. After achieving the best correction method, 10 other vegetation indices were calculated in addition to the two mentioned indices. Finally, after finding the best canopy estimation model, a vegetation canopy map was depicted for four time periods. Validation results showed that the FLAASH method is a superlative method in comparison to the other methods of QUAC, IRMAD and TOAin terms of RMSE and R. The R value was 0. 61, 0. 37, 0. 2 and 0. 57 for ARVI and 0. 54, 0. 39, 0. 21 and 0. 56 for NDVI, respectively. In addition, RMSE values were 0. 77, 0. 97, 1. 13 and 0. 8 for ARVI and 0. 83, 0. 96, 1. 12 and 0. 81 for NDVI, respectively. The vegetation canopy maps show the spatial heterogeneity of canopy cover in Marjan rangeland and its capability of estimating and monitoring the canopy cover of rangeland vegetation at different seasons using the developed model. canopy cover in Marjan rangeland and its capability of estimating and monitoring the canopy cover of rangeland vegetation at different seasons using the developed model.

Evaporation is one of the principal components of the water balance equation in a basin. For large water bodies located in arid regions, particularly in endorheic basins, evaporation loss is remarkable. Because of the considerable variation of hydro-meteorological variables over large basins, several models have been developed to estimate evapo-transpiration using remote sensing data. However, performance of such models over water bodies are questionable as they neither consider the lake geomorphology nor the water quality. Estimating evaporation from saline lakes is more difficult than that of the fresh water because the interaction of salinity and evaporation should also be considered. This study aims at comparing three energy balance models (EBMs) based on remote sensing data for estimating the rate of evaporation from Urmia Lake, one of the largest hypersaline lake worldwide. To do this, for 5 clear sky days in 2010, three commonly used RS-Based EBMs including SEBAL, SEBS and SDDE were applied using MODIS satellite imageries and results were compared against each other. Although all three models are capable of capturing spatial variations of meteorological variables over the lake, the only model which has been specially developed for water bodies and can completely reflect the effect of water salinity on evaporation is SDDE. The results indicate that the lowest lake-averaged evaporation rate is obtained in the SEBAL V. 1 and SEBAL V. 2 models, respectively. While the highest lake-averaged evaporation rates varies between the SEBS and SDDE models during the study days. Moreover, the evaporation rates from the lake decreases from the central deep parts toward the coasts in SEBAL and SDDE. This pattern is more homogeneous in SDDE. However, for the SEBS model, the spatial variation of the wind strongly depends on the spatial pattern of the wind speed. Comparison of the models with the pan evaporation data show that for most of the studied days, pan evaporation rates are less than the models estimations. Therefore, single-station pan evaporation measurements are inadequate to be used as representatives of evaporation from large lakes.

OSM is one of the most desired and recognized VGI projects. All information related to OSM data such as objects’ geometry, relations, and descriptive information including all previous versions is stored in the history file. Given the ease in access and the challenges in their quality debate, the issue of the quality of spatial information produced by OSM is an attractive topic for researchers. In previous researches, the use of the data history file to improve the quality of voluntary spatial information has not been considered. Hence, the goal of this research is to present a solution for improving the quality of location precision of the linear objects through the generation of new data using the data records. In this article, to achieve this goal, a geometric approach is presented based on the Voronoi diagram and object matching methods. To evaluate the effectiveness of the proposed method, the District 6 of Tehran was selected as the study area. In order to estimate the quality, the quality of extracted dataset was compared to Dataset produced by the municipality of Tehran as a reference dataset. According to the results obtained from this comparison, it was found that the completeness and positional accuracy of OSM features is improved by about 9. 03% and 8. 88%, respectively.

In recent years, proper management of resource and timing for emergency evacuation of citizens from the site of terrorist attack has been widely considered by researchers in the field of urban planning. Little studies which have carried out on modeling the behavior of citizens and terrorists in urban environment generally focused on terrorist’ s behavior. Although different age groups show different abilities in emergencies rised by terrorist attack, in those studies, citizens are modeled in only one group of age. The main purpose of this study is to develope an agent-based model to simulate emergency evacuation due to terrorist attacks considering available infrastructure of a city. In order to evaluate the potential terrorist attack in a city, the behaviour of three groups of agents, including civilians, defense forces, and terrorists, are developed and simulated. In this regard, by examining different scenarios, it is possible to minimize the casualties caused by terrorist attack and rapid evacuation. To perform sensibility analysis and determine the most sensitive parameters, One-at-a-time method is hereby employed. NetLogo software is also used to develop and implement the proposed agent-based model. As a case study, the presented agent-based model has been implemented under various scenarios in district 6 of Tehran and the results have been analyzed. Based on the results of analyzing difference scenarios, it is possible to minimize the possibility of terrorist attack by increasing the number of defense forces. Another useful scenario is to allocate 60% of the defense forces to protect the busy and safe areas and 40% to search for and stop terrorist agents in the study area, the number of terrorist attacks and the number of deaths will cut in half.

Distinctive and efficient description of image features is an essential task for image registration in photogrammetry and remote sensing. The majority of existing descriptors estimate a dominant orientation parameter for rotation invariant image matching. The dominant orientation assignment is an error-prone process, and it decreases the capability of the descriptors. In this paper, a novel feature descriptor based on the local binary pattern operator named RILBP (Rotation Invariant Local Binary Pattern) is proposed that is inherently rotation invariant. To compute the RILBP descriptor, the pixels in the given image region are divided into several sub-regions based on distance and intensity order constraints. Then, a local binary pattern histogram is generated for each sub-region based on a rotation invariant coordinate system. To increase the descriptor robustness against geometric distortions, a special weighting process based on a combined ring and Gaussian functions is applied. The proposed RILBP descriptor was successfully applied for matching of various remote sensing images as: SPOT 5, ETM+, Sentinel 2, IKONOS, IRS P6 and ZY3 sensors, and the results demonstrate its capability compared to common feature descriptors such as CS-LBP, SIFT, LSS, and MROGH. Compared to the standard CS-LBP descriptor, the RILBP descriptor indicates an average performance improvement of about 25%, 10% and 30%, in terms of Recall, Precision and number of correct matches, respectively.

The problem of extracting the building from mono optical aerial imagery with high spatial resolution is always considered as an important challenge to prepare the maps. The goal of the current research is to take advantage of the semantic segmentation of mono optical aerial imagery to extract the building which is realized based on the combination of deep convolutional neural networks (DCNN) and bilateral filters (BF). For this purpose, considering the hardware limitations of the current research and the fact that it is necessary to select a large number of training data to train deep convolutional neural networks, after selecting an appropriate dataset from three-band optical images, the minimum data that obtains the highest training accuracy was selected to avoid getting weak results due to the lack of training data. In this research, by optimizing the SegNet deep neural network which is an encoderdecoder network, the processing task and therefore extracting the building from optical images are done using the adaptive moment estimation (ADAM) optimization and BF with a Gaussian kernel. This method is implemented on a dataset related to the mono optical aerial imagery of urban regions located in Potsdam, Germany, the twodimensional tagged datasets of international society for photogrammetry and remote sensing (ISPRS). The results show that compared to similar methods, the combinational use of the SegNet optimized deep neural network and BF with a Gaussian kernel provides very appropriate capabilities to improve the detection of building boundary in the optical images with high spatial resolution. Also, the results of the proposed method show that the values of the integrity and validity criteria are 95. 14 and 92. 37 respectively for the test area 1, 91. 67 and 90. 2 respectively for the test area 2, and 96. 14 and 93. 98 respectively for the test area 3.