نشریه علوم و فنون نقشه برداری
سال هفتم شماره 1 (پیاپی 25، تابستان 1396)

Accurate positioning is an important problem in many fields especially Geographic Information System (GIS). With the advent of ubiquitous computing, a vast and massive change took place in various technologies, and a new generation of GIS, called ubiquitous GIS (UBGIS), was created. One of the most important aspects transformed by ubiquitous computing is in positioning process. ubiquitous GIS and its components provide an environment where position information can be accessed both inside and outside. Computer Vision and Vision based approaches could be a good and appropriate solution to improve positioning accuracy in a Ubiquitous GIS. Using simple and well-known objects such as targets is an appropriate method. Target recognition is important in determining the center and code of the target. In recent years, especial kind of targets called Coded Targets are considered in different fields of vision based approaches and the demand for a Coded Target guaranteeing automatic, error-free correspondence and accurate image point measurement, has been dramatically increased. Therefore, automatically detecting, matching, and determining the center coordinates of Coded Targets are critical issues. Due to various factors in the environment, automatic execution of this process is very difficult and complex or accuracy and speed are not suitable. This paper aims to propose a new method using feature based matching algorithms for automatically recognizing Coded Targets and identifying their centers with sub-pixel accuracy, which can be used to enhance positioning accuracy in ubiquitous GIS. To achieve this aim, feature based matching algorithms and combining local feature detectors and descriptors like SIFT, SURF, and AKAZE are used to find corresponding image points and automatically recognition of Coded Targets. Therefore, suitable matching algorithm is chosen by comparing different matching algorithms. Results show that the best matching algorithm for this usage is SIFT-SURF that means using SIFT descriptors and SURF detectors will lead to best matching results. Then K-means clustering method is applied to distinguish Coded Targets and extract code of target with respect to template targets that are stored in the database. The cluster with the largest number of corresponding points belongs to the template Coded Target. In second stage a bounding box around the matched Coded Target is considered by defining minimum and maximum coordinates of corresponding points around target, so that there is only one target in this boundary. Then the image is cut in this boundary to firstly increase the speed of the calculation, because the search area for an image gets smaller and secondly reduce the possibility of mistake, because the other features and targets in the cut image are almost eliminated. Then center coordinates of Coded Targets are computed by finding contours in this bounding box and fitting a Hough ellipse to central ellipse of target. Finally, the center of this fitted ellipse is computed as the center of Coded Target. The results of implementing the methods are compared with well-known photogrammetry software called Agisoft (modelling and accurate measuring based on basic photogrammetry and computer vision). Results demonstrate sub pixel accuracy {0.574, 0.496 pixel} in center determination in X and Y direction respectively and success possibility of 63% in code recognition.

Identification people's regions of interest and ranking them based on their popularity level have many applications such as tourism development, traffic management and urban planning. Most of interesting places from tourism's view-points are parks, museums, historical places and scenic areas. To identify these places, it is necessary to access people's perception about environment. Accessing to this environmental perception isnt possible from traditional maps or satellite images. One approach to achieve this kind of people's environmental perception is VEP[1]. In this method, people are asked to take photos about subject of research and then the content of these photos are analyzed. So in the VEP method, the content of images is very important part. However, todays with the development of web 2, people share their taken photos from scenic places and interesting subjects. Each photo has metadata about its spatial location, identity and name of its up-loader, title and more. Therefore, researches in the field of geo-tagged photos, analyzed metadata of these photos. However, in this research, we investigated both of photo's content and metadata to rank popular places and it can be said that the integration of VEP and VGI methods is our main contribution.
Tehran region 6 was considered as the case study and its geo-tagged photos are extracted from Panoramio site. Then DBSCAN[2] method was applied to extract regions of interest. The DBSCAN method has many advantages that are 1) it is density-base and the place that have dense data points is identified as cluster. Therefore, it is appropriate for our research that density of photos is an indicator of place's popularity 2) Moreover, in this method, there is no need to know the number of clusters previously and 3) the shape of clusters can be arbitrary.
Two mandatory parameters of the DBSCAN method are Eps (Neighborhood’s radius) and MinPts (at-least number of points) but there isnt any ideal method to obtain optimal values of these parameters for all applications. Therefore, to find appropriate values, we ran DBSCAN with different parameters and finally we set Eps as 1000 and MinPts as 10, and in result, 17 clusters were identified. The concept of each cluster was identified based on GeoNames POI[3] and unrelated clusters of tourism were removed. Then the popularity score of each region of interest was computed based on its photo's contents, number of photos and number of up-loaders. One of the scene-recognition algorithms was applied to investigate photo's content. The Laleh Park, Saei Park and Ferdosi square achieved high popularity scores. In the next step, popularity of these places in different months, seasons and years were investigated. Totally it can be said that most of the photos were taken in the April and May, or in other words, in the spring. Moreover, the relation between regions of interest and their land-use types were investigated that shows that green-space was significantly more than other land- use types. The detected places were compared with tourist attractions in Tehran region 6 and this comparison showed that natural attractiveness such as parks and gardens have appeared more than other attractiveness in geo-tagged photos. Comparing the computed popularity score of each region of interest, with sum of its scores on GoogleMap and FourSquare, showed that more coincidence exists in the class of very popular places.
[1] Volunteer-Employed Photography
[2] Density Based Spatial Clustering With Noise
[3] Points of Interest

Cities are growing and encountering many changes over time due to population growth and migration. Identification and detection of these changes play important roles in urban management and sustainable development. Urban growth models are divided into two main categories: first cellular models which are further divided into experimental, dynamic, and integrated models and second vector models. In this study, an integrated urban growth model is proposed which is a combination of cellular automata and gravitational search algorithm (GSA). It has been implemented on Shiraz (Iran) to model the urban growth between 1990 and 2000. The proposed integrated model uses GSA to calibrate cellular automata transition rules. The Landsat satellite imageries in 1990 and 2000 with Digital Elevation Model (DEM) of Shiraz are used in this study. Five parameters including distances from major roads, urban neighborhood, slope, distances from attraction centers, and distances from parks and other green spaces are considered to be effective in the urban growth modeling. Based on the results, Kappa coefficient and overall accuracy of the model are 66.54% and 92%, respectively. By using GSA, calibration of cellular automata is facilitated and the proposed integrated model reaches optimal solutions in fewer iterations. The achieved results show that the proposed integrated model can be used for studying urban growth.

The ever-increasing population in cities intensifies environmental pollution that increases the number of asthmatic patients. Other factors that may influence the prevalence of asthma are atmospheric parameters, physiographic elements and personal characteristics. These affecting parameters can be incorporated into a model to monitor and predict the health conditions of asthmatic patients in various contexts. Such a model is the base for any asthma early warning system.
With the rapid advancement of human knowledge in diverse areas, new science and technology has been offered to aid people in terms of education, food, transportation and health. Ubiquitous computing is one of the newest human developments to enhance individuals’ lives. In recent years, the efficiency of ubiquitous computing in a wide range of applications such as government, health, safety, municipal and transportation have been studied and validated.
This paper introduces a novel ubiquitous health system to monitor asthmatic patients. Ubiquitous systems can be effective in monitoring asthmatic patients through the use of intelligent frameworks. Our paper proposed a model for prediction of asthma conditions in various scenarios. The asthmatic conditions of patients were predicted accurately by a Graph-Based Support Vector Machine (SVM) which functions anywhere, anytime and with any status. Proposed model is an improved version of the common SVM algorithm with the addition of unlabeled data and graph-based rules in a context space.
The study graph was formed by the :union: of the training data (L) and the unlabeled data (U). Afterward, the best kernel type for SVM was estimated, and a multi-class SVM algorithm was performed. Initial classification was carried out and the U dataset was tagged. Next the k-nearest neighbor was determined around each training data item followed by the weighing of each edge of the graph (Wij) based on inverse Euclidean distance. This implies that larger weights were given to any unlabeled data close to the training data. Unlabeled data with high weights were assigned the same label as the reference training data and then the position of the data was varied. In this article, “position” means the location of data point in context space. The positional change was performed to closing the unlabeled data to the training data with respect to its weight. Then, the context space was updated and the SVM algorithm was run. This process continued until an acceptable threshold () was reached. At the end of the process, the final labels were assigned to unlabeled data and the PEF conditions of each patient were predicted.
Based on the stored value for a patient's condition and his/her location/time, asthmatic patients can be monitored and appropriate alerts will be given. Our proposed model was assessed in Region 3 of the city of Tehran, Iran for monitoring 3 different types of asthma. The input data to our asthma monitoring system included air pollution data, the patients’ personal information, patients’ locations, weather data and geographical information for 270 different situations. Our results ascertained that 90% of the system’s predictions were correct. The proposed model also improved the estimation accuracy by 12% in comparison to SVM and ACO methods.

The study of surface displacement, particularly subsidence occurred in important ground and underground structures, and identification of surface dynamic behavior of these structures are always one of the major challenges in recent centuries that have too much costs for various organization. So far, there are several tools to measure the surface displacement, among them; one can point to Global Positioning System (GPS) and precision leveling. These tools measures the surface displacement of the earth at large scales (spatial and temporal) with high precision. However, despite the advantages of this method, these tools often very costly, time consuming and requires intensive field work, and additional this method carried out over a small area and only on a limited number of target points or on a single line which passes through the area. Therefore, do not extract the full details of the deformation field in space and time. These reasons have caused attention to a method that reduces these limitations. In the resent year one of the newly technique for detected this displacement and subsidence is the Interferometry Synthetic Aperture Radar (InSAR), persistent/permanent scatterer interferometry (PSInSAR) was developed to improve the accuracy and precision of the deformation measurements, even reach to sub-millimeter level. They are several algorithms for deformation monitoring using PSInSAR, this methods uses a huge collection of stable pixels (PSs) for processing which are known as coherent targets how stable the phase over the time. Use of PSs points causes further increasing the spatial density of deformation-tracking points or improve the accuracy of deformation measurements, therefore in this study used of the persistent/permanent scatterer Interferometry Synthetic Aperture Radar (PSInSAR) and height resolution radar image in order to investigation the subsidence of structures and important ground and underground targets in Tehran city between 2014 to 2016. Our radar data set consists of 13 StripMap HIMAGE, right-looking, ascending mode scenes with an image swath of 40 km, which were acquired at 3-m resolution and with HH polarization, between 2014 to 2016, from the SAR sensors on-board the four satellites of the COSMO-SkyMed constellation, at a ~620-km altitude above the Earth’s surface with an LOS incidence angle of 36° at the scene centre, a minimum of 32.4° and a maximum of 35.5° at the near and far range, respectively, In present study, StaMPS persistent scatterer approach, under SBAS and PS algorithms is used for investigation of time series Interferometric analysis. The results of both processing indicated that significant subsidence in southwest part of Tehran city and in 9, 17, 18, 19 and 21 districts of Tehran municipality during this two year, is occurred which average rate of it is around of 61 mm/year. Furthermore, in this study, displacements occurred in the location of some important structures including Milad tower, Azadi stadium, Azadi tower and Mehrabad airport and also surface displacement enroute Tehran subway, based on extracted results from time series radar Interferometric processing and detecting of danger regions, is evaluated. As a result of this evaluation, the regions that are subject of danger (regions exposed to danger), would be identified and introduced to related organizations for suitable prevention function in future.

Search, discovery, and prioritizing important places and facilities is a vital requirement of users in urban environments. Over the past twenty years, web mapping applications have been continuously improving to better proffer location based services to users. A considerable part of users’ search in web mapping application can be addressed by simple distance-based spatial search. However, there are cases where, in order to properly address user requirements, web mapping applications need to consider not only simple spatial criteria but also spatial relationships. Additionally, current web is moving towards machine readable and machine understandable contents through semantic web. Semantic web provides standards and technologies for machines to be capable of understanding documents on the web. Meanwhile, the amount of spatial data that is published on the web is rapidly growing. These data are faced with problems such as diversity, heterogeneity, incompleteness, and therefor are not well prepared for retrieval and utilization. Description of spatial data using semantic web concepts is a solution to most of the mentioned problems. Currently, different standardization organizations in geospatial community are working on development of required standards for semantic spatial data. GeoSPARQL, a semantic query language, is a standard developed to query and retrieve semantically described spatial information over the web. In order to enable citizens to search and recommend urban facilities, this study developed a web mapping application. In this application, GIS and semantic web technologies are integrated to enhance the performance of the system. Search and discovery based on simple geometric relations, topological relationships, spatial operators, semantic relations and descriptive characteristics of features are implemented in the system. The system implements semantic web technologies, ontology, Open World Assumption and spatio-semantic query language. GeoSPARQL ontology is used to define a RDF/OWL vocabulary for representing spatial objects and querying topological and non-topological relations. QallMe ontology is used to define concepts related to the tourism industry including urban facilities. In this system, interests and needs of citizens are considered as well. There are two executive routines in the system, including data entry and query handling. In the data preparation phase, spatial characteristics and attribute information of urban service centers from various sources are gathered and converted to a semantic representation based on the application ontologies. This semantic representation is saved in Parliament triple store. Having the data prepared in the triple store, a user can access the system through a web based user interface and search for required facilities or places within the city. The query is transformed to GeoSPARQL Select queries afterward and executed on the triple store. The responses are presented to the user on a map, developed by Google Maps APIs. In order to demonstrate the feasibility of the developed application, two usage scenarios are thoroughly implemented and described. The implementation results in spatial software comparing the developed application show the general adequacy of the system. Additionally, diversity, heterogeneity and incompleteness problems of spatial information on the web have decreased to a significant extent. The developed system helps citizens to find urban facilities accurately and efficiently.

Nowadays, severe traffic congestion in urban areas resulting in different undesirable socio-economic and environmental consequences is inevitable. Infrastructure improvement for preventing these undesirable impacts seems to be necessary. Integration of intelligent transportation systems (ITS) into the existing transportation infrastructure leads efficient operation -using electronic, sensing, information and communication technologies, and advanced control techniques- without building new roads.
The main focus of this article is developing multi-agent reinforcement learning for traffic signal control. Two types of agents are employed: (1) Learning traffic signal agents (LTSAs) that interact with the traffic environment in order to find the optimal traffic signal parameters (traffic signal timing) in response to traffic fluctuations. (2) Vehicle agents that are purely reactive. They can detect their forward direction, current driving lane, other vehicles, and the current phase of approaching traffic signal. Also, vehicles can chane their driving lane in order to reach the better driving speed. Unlike vehicles that are reactive and are not able to learn, LTSAs have the ability to learn over time through reinforcement learning.
Reinforcement learning originally stems from the study of animal intelligence and has been developed as a major branch of machine learning for solving sequential decision-making problems. It is a useful approach for solving the stochastic optimization problems. It learns the optimal policy of the agent by interacting with the environment in such a way to maximize some numerical value which represents a long-term objective. Reinforcement learning allows traffic signals to automatically determine the ideal behavior for achieving their objectives. In fact, it enables traffic signals to learn and react flexibly to different traffic situations without the need of a predefined model of the environment and also without the need of human intervention. Each time the traffic signal performs an action, it receives a reward signal indicating whether its action has led it closer to realizing their objectives or not. The traffic signal tries to learn a control policy which is a mapping from states to actions that maximizes the expected sum of the received rewards.
Two different scenario including single-agent traffic signal control and multi-agent traffic signal control were conducted. In the first scenario, a learning agent controls an isolated intersection by employing two methods of reinforcement learning including Q-learning and State-Action-Reward-State-Action (SARSA). Q-Learning is an off-policy method that updates the value of actions based on the hypothetical actions. In Q-Learning, as long as the traffic signal visits all the state-action pairs, it converges to the optimal action-values. SARSA is an on-policy algorithm that updates action-values on the basis of the experience gained from following some policy. In SARSA, the traffic signal should explore, and stop exploring after a number of steps. The results of the first scenario indicate that Q-Learning outperforms SARSA. In the second scenario, four learning agents control the main street composed of four intersections by employing indirect cooperative Q-Learning. The results of the second scenario reveal that the indirect cooperative Q-Learning controller decreases 81% queue length, 78% travel time, 57% fuel consumption and 73% air pollution when compared to the optimized pre-timed controller.

Ionosphere is a layer in the upper part of atmosphere wide-ranging from 60 km to 2000 km. It has a very significance in radio wave propagation because of, its electromagnetic attributes. Ionosphere is mainly affected by solar zenith angle and solar activity. In the day-time ionization in ionosphere is at the highest level and the ionospheric effects are stronger. In the night-time ionization decreases and the effects of ionosphere gets weaker. One of the most important parameters that define the physical structure of ionosphere is total electron content (TEC). TEC is a line integral of electron density along signal path between satellites to the receiver on the ground. The unit of TEC is TECU and 1 TECU equals 1016 electrons/m2. The TEC values can be computed from dual frequency global positioning system (GPS), which are the most available observations for studying the earth’s ionosphere. However, because of scatter repartition of dual frequency GPS stations, precise information on TEC over the favorable region is unknown.
Artificial neural networks appeared in the 1980 of the 20th century, it uses physical systems which can be realized to simulate the human brain structure and function of nerve cells. With distributed storage, parallel processing, the ANN has good self-earning, adaptive and associative function, can adapt to the complex and ever-hanging dynamics characteristics. Figure 1 shows the scheme of a three-layer perceptron network. For training of the network and modifications of the weights, there are so many ways. One of the most famous and simplest methods is back-propagation algorithm which trains network in two stages: feed-forward and feed-backward. In feed-forward process, input parameters move to output layer. In this stage, output parameters are compared with known parameters and the errors is identified. The next stage is done feed-backward. In this stage, the errors move from output layer to input layer. Again, the input weights are calculated. These two stages are repeated until the errors reaches a threshold expected for output parameters.
Particle swarm optimization (PSO) is a population based (evolutionary) stochastic optimization technique in which a collection of particles move around in search of space looking for the best solution to an optimization problem. The concept is derived from the motion of a flock of birds that communicates and learns from each other in search for food. This algorithm proposed by Eberhart et al., (2001). A PSO algorithm is inspired on the movements of the best member of the population and at the same time also on their own experience. The metaphor indicates that a set of solutions is moving in a search space with the aim to achieve the best position or solution.
In this paper, 3-layer MLP-ANN with 18 neurons in hidden layer is used to modeling the ionosphere TEC time series variations. For this purpose, observations from 36 GPS station in 11 successive days of 2012 (DOY# 220 to 230) are used to processes. To accelerate training step and also enhance the accuracy of the results, particle swarm optimization (PSO) algorithm is used. GPS TEC is used to validate the accuracy of results. Also results of ANN compared with international reference ionosphere (IRI-2012) and universal Kriging method. Analysis of the results showed that the PSO training algorithm has a high-speed in convergence to the optimal solutions. To evaluate the error of ANN results, dVTEC=VTECGPS - VTECM is used. Minimum dVTEC is computed 0.55, 1.57 and 0.70 TECU for ANN, IRI-2012 and universal kriging methods. Also, maximum dVTEC obtained 5.45, 7.16 and 5.51 TECU, respectively. The results of this paper suggest that the artificial neural network with PSO training algorithm has high accuracy in modeling of ionosphere electron content time series.

In this study, the problem of local gravity field modeling is investigated with thw aid of spherical cap harmonic analysis. To do so, the Dirikhlet boundary value problem for Laplace equation is solved for boundary condition prescribed on the surface of a spherical cap. In this case, the solution of Laplace equation is represented based on linear combinations of associated Legendre function of non-integer degree and integer order. To evaluate the performance of the model, a spherical cap zone located in northwest of Iran with a half-angle of one degree is selected and the local gravity data in the form airborn observations are simulated over the cap region using global geopotential model. Morover the simulated data are contaminated with random noise in order to better adapt with actual airborn observations. Using thses observation in series representation by pherical cap harmonic analysis, then, the geopotential coefficients for local gravity field are computed. Since the govering equations for determination of the cofficients suffer from an ill-condition problem, it is necessary that some regularization schemes are applied. Here, the Tikhonov regularization method is utilized to obtaine the regular solution. To validate the accuracy of proposed model over the cap region, results are compared with observation of gravity at some control points distribute both within the cap and on its margin.

Rapid urbanization and urban growth are important issues to be considered. Urban development factors recognition can help urban planners and decision makers to understand the consequences of their decisions on urban growth and development .So urban planners, in order to Control the urban growth, use dynamic systems .cellular automata (CA) models have been increasingly used over the last decades to simulate a wide range of spatial phenomena.Urban development is a complex spatio-temporal process that involves both horizontal and vertical growth. Despite growing recognition of the significance of horizontal development, models of urban vertical growth remain limited .This study aims to develop a GIS-based vector cellular automata model for exploring the Horizontal growth and vertical complexities of urban growth according to the height states of existing buildings in the neighborhood and using General Development Control Regulations.The developed model has the capacity to simulate urban growth space and hence vertical growth .series of variables that are used in horizontal model including accessibility, physical suitibility , neighborhood affect and land price.Affactive factors of urban development are classified in two part: vertical variables and horizontal variables .Slope and Height factors will be used in order to calculate the phizical suitability for undevelop parcel. Euclidean distance and network distance will be used in order to calculate accessibility. neighborhood affect is considered in three component. Also Value of residential, office and retail floor area are computed for property values affect.
Vertical model was performed in two different ways. In the first method, undevelop Parcel is assigned a state value according to the height states of existing buildings in the neighborhood in which it is situated. In the second method, using General Development Control Regulations can simulate different height states of building growth. the model has been developed in a very small area of ​​the city of Qom in the years 1385 to 1394 .The results show overall accuracy 60% in horizontal growth and overall accuracy 65 % in the first method of vertical growth and overall accuracy 60 % in the second method of vertical growth .Therefore, it carries scope of being used to visualize growth for other, similar, cities and help urban planners and decision makers to understand the consequences of their decisions on urban growth and development.

Almost all MODIS images have cloudy regions. Cloud detection and discriminating it from similar objects like snow/ice is a necessary pre-processing step before extracting accurate information. Features for cloud detection can be divided into two categories; textural and spectral features. Using textural features in the visible bands make cloud and snow/ice pixels separable, while cloud and snow can have similar temperature. Thermal features have been used for cloud detection in different heights in MODIS cloud mask (MOD35). Recently variety of studies using ensemble learning methods for land cover classification have been done. In the studies, ensemble learning methods used for classification but in this study, a new application has been introduced; for fusion of classifiers having reflectance and thermal features. Also, in previous studies, the effect of changing feature selection method on the performance of ensemble learning methods has not been examined. So, the purpose of this study is the comparison of the performance of fusion of the reflectance and thermal classifiers using two kinds of ensemble learning methods including boosting and Random Forest (RF) for detection of cloud, cirrus and snow/ice pixels, based on the feature selection method applied. First, some of the Visible-Infrared bands (VIR), 1, 2, 8 and 26 in addition to the thermal bands including 20, 22, 31, 32 and 35 of Terra MODIS were calibrated and the reflectance and Brightness Temperature (BT) values were extracted. Also, three indexes, Normalized Difference Vegetation Index (NDVI), Normalized Difference Snow Index (NDSI) and the ratio of bands b1/b2 were computed. Gray Level Co-occurrence Matrix (GLCM) textural features of the above mentioned bands and some of the BT differences including, BT31-BT32, BT33-BT31, BT31-BT20, BT20-BT32, BT22-BT32 were added to the reflectance and thermal input features sets. For selecting the suitable reflectance and thermal features in different kinds of boosting methods including Adaboost.M1, AdaboostSVM, Logitboost and Totalboost, S criteria and Genetic Algorithm (GA) were used, and in RF algorithm in addition to these methods, Recursive Feature Elimination (RFE) and correlation matrix were applied too. After feature selection step, training data were selected manually for cloud, cirrus and snow/ice and fed into the reflectance and thermal classifiers. Classifiers were fused in decision level using the majority vote method. The performance of different cases was compared using producer accuracy, user accuracy and kappa coefficient indexes. For almost all ensemble methods, ignoring which one of the feature selection methods applied, a high cloud and cirrus producer accuracy achieved. RFE and correlation matrix methods in RF algorithm, resulted in 99% and 100% cloud producer accuracy. These values are higher than S criteria and GA. Boosting methods, ignoring the kind of feature selection method, got higher snow/ice producer accuracy than RF algorithm by assigning higher weights to the class that has less training data among classes. Also, boosting methods resulted in higher cirrus user accuracy than RF. Among feature selection methods applied in RF, correlation matrix achieved 91% value for cirrus user accuracy. Finally, agreement to the reference map produced from MOD35 calculated. RF algorithms showed higher agreement with the reference map in comparison to the boosting methods. Highest agreement resulted from RF-RFE with the value of 76% and lowest agreement from Logitboost-GA with the value of 42%.

The Middle East is one of the hot-spot regions that has reported the highest occurrence of dust storms. In this study changes in aerosol optical and radiative properties during dust storm over the Middle East on July 2009, was analyzed using retrieved Aerosol Optical Depth (AOD) from Simplified Aerosol Retrieval Algorithm (SARA) algorithm over study regions and ground-based observations from CIMEL Sunphotometer measurements at AERONET (AEronet RObotic NETwork AERONET) site at Kuwait University. Dust plumes captured well with the SARA AOD during dusty days. Maximum AOD values occurred on the 4 July in Kuwait, Ahvaz, Iran, the Persian Gulf, Baghdad and Riyadh where values of 1.44, 2.56, 1.07, 3.0, and 1.99 were recorded, respectively. The maximum aerosol volume size distributions (VSDs) at AERONET site occurred on dusty days and minimum VSDs on non-dusty days. The single scattering albedo (SSA) obtained higher values on dusty days compared to non-dusty days, for all wavelengths (440, 675, 870, and 1020 nm) at the AERONET site. The maximum SSA value of 0.99 occurred at a wavelength of 1020 nm, which is an indication of the dust aerosol. The asymmetry parameter (AP) obtained higher values at shorter wavelengths over the study period. The AP (AP-T) for both modes (fine and coarse) values were higher in the near infrared region than in the visible spectral region on both dusty and non-dusty days. The coarse AP (AP-C) values were higher in the visible spectral region than in the near infrared region on dusty days with the reverse being true on non-dusty days. The VSD, SSA, AP, and refractive index values on dusty days suggested that dust aerosols were predominant over anthropogenic aerosols in the study area. Effect of aerosol on Downward Surface Shortwave Radiation (ARFDSSR) on both dusty and non-dusty days ranged between -51 and -160 Wm−2 (average: -90 Wm−2) at the earth's surface. Effect of aerosol on Net Surface Shortwave Radiation (ARFNSSR) on both dusty and non-dusty days ranged between -39 and -122 Wm−2 (average: -69 Wm−2) at the earth's surface. Dust-induced turbid conditions caused significant extinction of 15–20%, in DSSR, resulting in an increase of 66% in aerosol radiative forcing (ARF) at the AERONET site at Kuwait University. The HYSPLIT back-trajectory analysis revealed that the air masses reached Kuwait from the western part of the Sahara Desert in northwest Africa and Saudi Arabia over the Middle East. Aerosol vertical profiles retrieved by the space-borne Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) reveal a well-mixed dust layer occurred over the study area. The aerosol types identified by CALIPSO in the near of the study regions include both dust and polluted dust, but dust aerosols were dominant over the anthropogenic (polluted) aerosols. The CALIPSO aerosol profile indicated a layer of thick dust extending from the surface to an altitude of about 6 km. The results of radiative forcing of dust storm in the Middle East show that mineral dust cause decrease of shortwave radiation and net radiation at the surface. Additionally the presence of mineral dust cause decrease of temperature at earth’s surface. So, dust directly influence the earth’s radiative budget and cause surface cooling.

Unmanned Aerial Vehicles (UAVs) are promising tools for many applications, including agriculture, mining, recreation, search and rescue, infrastructure monitoring, and wildlife research and conservation. Many of these applications require some type of object tracking. In fact, one of the most valuable capabilities of UAVs is their ability to detect and track particular moving targets on the ground. The main problem in moving target tracking using UAVs is the background movement with the intended target. In this study, a new method for detecting and tracking moving targets using frame-to-frame registration is presented. The proposed method does not require other external sources to know about the position and orientation of the platform. Also, computer vision processes are applied to detect moving objects and the detections are passed to a tracking algorithm, which then generates continuous tracks of objects seen by the camera.
This paper first reviews the current methods utilized in multiple target tracking in video with particular emphasis on airborne applications. It then presents a method for detecting moving objects in video obtained from a UAV. In this method, the improved Speeded-Up Robust Features (SURF) matching algorithm that adds color information to conventional SURF descriptor is used to enhance matching results. The SURF algorithm is widely applied in the field of target detection and tracking. But in this study, the advantages of SURF and color information are combined to achieve better results. Then, Random Sample Consensus (RANSAC) technique is used to remove weak matches. After finding reliable corresponding points in consecutive frame images, the projective transformation between two frames is calculated. Finally for detection the moving object, frame subtraction and image segmentation methods are used respectively. Generally, the proposed method can be summarized like this: the suggested matching method using the improved SURF matching algorithm.
The proposed method is implemented on video data obtained from a UAV which has a HD camera with the frame rate of 60 frames per second (fps) and the resolution of 1080×1920 pixels. It is studied to detect moving vehicles and people and the derived outcomes are analyzed. Experimental results show that the proposed method has suitable performance for moving object detection in images with moving background. However, the results show that although the improved SURF matching algorithm imposes some computational complexities to the proposed method, it eventuates to positive effects in matching results.
Average processing time of the proposed method is also studied. Results show that although the average processing time is increased about 6%, the matching accuracy increased by an average of 8.5%. However, due to the important role of the matching accuracy on the accuracy of final results obtained from proposed algorithm, using the improved SURF matching algorithm is an appropriate choice. Also, other deliberations show efficiency of the proposed method in detection and tracking moving objects.
It is also suggested that in segmentation stage, the performance of other techniques like Fuzzy segmentation algorithm would be evaluated in comparison with Niblack technique. Also the gained results of the proposed algorithm would be evaluated on the data recorded in different environmental conditions like different lightening situation, excessive shaking of the camera and foggy and dusty weather.

Thermal infrared bands contain important information for various applications. Their spatial resolution is relatively low, and it is hard to determining the location of the targets in these bands. The aim of this study is to enhance spatial resolution of thermal bands. Image fusion is one of the efficient methods that are employed to enhance spatial resolution of thermal bands by fusing these data with high spatial resolution visible bands. Image fusion aims to integrate images with different spatial and spectral resolution, such that the synthesized image is more suitable for human visual perception or further processing, such as image classification, segmentation, texture feature extraction, object recognition, etc. We chose pixel level image fusion and added details and spatial information taken from the visible band to the thermal infrared band. Multi-resolution analysis (e.g. wavelet, laplacian pyramid, contourlet, curvelet, etc) is an effective pixel level fusion approach. In this paper, the contourlet transform in image fusion due to its advantages, high directionality and anisotropy is used. Because of the downsampling and upsampling, the contourlet transform lacks of shift invariance and results in artifacts. Therefore, we use the other kinds of contourlet transform, nonsubsampled and sharp frequency localization contourlet transform, and then, the image fusion performance of six multi-resolution transforms, including the discrete wavelet, stationary wavelet, laplacian pyramid, contourlet, nonsubsampled and sharp frequency localization contourlet transform are compared. The methods have been tested using thermal infrared and visible landsat-8 data. The spectral and spatial quality assessment parameters (e.g. CC, SAM, ERGAS, SNR, RMSE, AG, UIQI, etc) show that the sharp frequency localization and the nonsubsampled contourlet transforms perform better than the discrete wavelet and the original contourlet transforms in terms of preserving radiance spectral information and increasing spatial details The experimental results show that the discrete wavelet and the contourlet transform are the worse than the other transformations. Therefore, the shift invariant property is of great importance for image fusion. The results of our comparative analysis show that in spite of the lack of a spectral overlap between the visible and the thermal infrared bands, the final fused thermal image keep its spectral characteristics while the spatial resolution is enhanced. It can be concluded that the sharp frequency localization contourlet transform with redundancy factor of 2.33 is the best technique for fusion of the visible and the thermal infrared bands. Finally, the effect of decomposition levels, on fusion performance result by sharp frequency localization contourlet transform with redundancy factor of 2.33 was investigated. It also can be concluded the appropriate setting for the number of decomposition levels is four.

Nowadays spatial data and urban areas rapidly changing due to the many kinds of natural and artificial factors. These changes lead to the loss of reliability of the information for urban planning, resource management and inefficiency of spatial information systems. so, monitoring of these changes and obtaining update information about the land use and the kind of its changes is essential for urban planning, proper resource management, damage determination assessment and the updating of geospatial information systems. Therefore, more accurate change detection is a challenge for experts and researchers of remote sensing and photogrammetry. In recent years, various techniques have been developed for change detection especially on high-resolution images that choosing the appropriate method and algorithm to identify changes is not easy. Despite all the efforts of researchers to develop different methods for change detection, all techniques and methods have advantages and limitations. This article introduces a new category of changes detection methods. In general, methods and techniques of change detection in urban areas can be categorized into four major categories: direct comparison- post classification dipole, object based- pixel based dipole, supervised- unsupervised dipole, textural and spatial information and features. Despite a rich and useful spectral information in high-resolution satellite images of remote sensing and photogrammetry, just use of this kind of information, will not be enough to achieve the required accuracy due to increased variability within homogenous land-cover classes. So, in this paper, in addition to the spectral features, it is also used texture features extracted from the spatial and frequency domain (Spectral, Anomaly, Edge, Morphological building index (MBI), Other color space, Gray Level Co-occurrence Matrix (GLCM), Features extracted from wavelet transform, Features extracted from Gabor filter, Features extracted from Fourier transform and Features extracted from curvelet transform) to solving this problem and generating changes mask of high-resolution images. The diversity and variety of extracted features from the spatial and frequency domain require optimization algorithms to achieve optimum features. Therefore, particle swarm optimization and genetic algorithms have been used to achieve optimum features and optimum parameters of support vector machine simultaneously. Also according to the major weakness of post classification method for detection of intra-class changes and bad radiometric conditions of used images for segmentation, 2-class classification of differential features is used to detect changes. QuickBird (0.6 m - October 2006) and GeoEye (0.5 m - August 2010) satellite imagery of AzadShahr/Tehran/Iran are used to evaluate the proposed method. The overall accuracy 93.45 and kappa coefficient 0.87 versus 91.03 and 0.82 show that particle swarm optimization is better than a genetic algorithm to achieve optimum features and optimum parameters of support vector machine simultaneously. It also calculates the effectiveness of each 10 kinds of features used by three criteria introduced in this paper (Effectiveness, Minor Effectiveness, and Overall Effectiveness), indicates the efficiency of using other color spaces, features extracted from wavelet and features extracted from spatial domain (Gray Level Co-occurrence Matrix) and also reflects the weakness of using only spectral data to detect changes in high-resolution images. Compare the proposed approach with other studies (post classification and fuzzy thresholding method) show the effectiveness of proposed method.

Satellite image time-series (SITS) data are a set of satellite images acquired from the same geographic area over a period of time. SITS data, due to their ability to capture the dynamic spectral behavior of the crop during their growing season, have been increasingly used for accurate crop mapping. The time-series obtained from the multispectral or hyperspectral sensors can be considered as multivariate time-series. The classification of these type of time-series data is a challenging task. This is mainly due to the fact that these data can be considered as a four-dimensional data, so the available classification algorithms cannot be used for their classification. To address this issue, in this paper by using the Multiple Kernel Learning (MKL) algorithms, a novel method for classification of multivariate time-series data is proposed. MKL algorithms are a group of kernel learning algorithms that make it possible to use a combination of kernels instead of a single one for kernel-based learning algorithm such as classification. In the proposed method, initially one kernel is constructed from data of each time of the time-series and then by using the MKL algorithms, these kernels are optimally combined into a composite kernel. The obtained composite kernel, once constructed, can be used to classify the time-series data by using all the kernel-based classification algorithms. In order to evaluate the proposed method, two time-series data were used. Both these time-series consisted of 10 different RapidEye imageries, acquired over an agricultural area in Manitoba, Canada. Both these time-series contained the main crop types of the region such as wheat, corn, canola, and soybean. In order to evaluate the effects of different MKL algorithms in the framework of the proposed method, in addition to the common MKL algorithms, the Generalized Multiple Kernel Learning algorithm (GMKL) was adopted as the MKL algorithms in the proposed Method. The GMKL is one of the most recent MKL algorithms proposed in machine learning literature, which has not been evaluated for the time-series data analyses. As a benchmark for comparison with a single kernel method, stacking method, in which the data acquired at different times are stacked into a single data cube, was used. The composite kernel obtained from the proposed algorithm with adopting different MKL algorithms and the kernel constructed from the data cube obtained from the stacking of the data were used to train of a Support vector machine algorithm. The obtained classification accuracies of the SVM showed a dramatic increase (at least 4.9% increase in the overall accuracy of the classification) in the case of using the kernel obtained from the proposed method in comparison with the case of using the kernel obtained from the stacked data cube. Moreover, the GMKL algorithm showed a higher performance in comparison to other MKL algorithms in the context of the proposed method for multivariate time-series classification. In addition, the proposed method showed better performance in the presence of cloud and cloud shadows in the data. This is because the MKL algorithms can reduce the negative effects of the cloud contaminated images within the time-series.