فصلنامه سنجش از دور و GIS ایران
سال نهم شماره 3 (پیاپی 35، پاییز 1396)

This research studies the suitable process of change detection at at an Agricultural areas by focusing on object based method and color fusion. In order to obtain this goal, it is benefit from Landsat7 images. The main idea of offering object based method is a modern algorithm i.e. Double-layer image are combined and An image of the entire layer is formed. Then by selecting suitable parameters a single image is separated in to several parts and by color fusion and object based classification method the changed and unchanged parts are classified. In fact, color fusion is determined by creating different color areas with elementary images that determines changed parts on visual basics and then by using object based classification method and selecting some parts by the user, the total parts of image is determined. Finally, by selecting training samples only one part of image is labeled and its classification is determined and the ultimate map of changes is obtained. Results show that this method is suitable for reducing training samples, increasing exactness (3%-2.5%), speed and increasing information for classification of spatial information and structure and in addition to spectral information it is better than ordinary methods of change detection from comparing 2 multi-temporal images.

Today, one of the limitations of water resources is the solutions weakness of water resource management. One of the management solutions to improve the problem above is the optimal allocation of users with the virtual water approach. In the present study, a model for optimization of user allocation with a virtual water storage approach is provided with using of genetic algorithms (NSGA-II, GA) in HAJILAK lands, located in the city of BUKAN in the West AZARBAIJAN province. After the preparation of the land use layer in the GIS and the preparation of target function coefficients, the user allocation with using meta-heuristic algorithms, is optimized with particular attention to virtual water. The results show that the suggested user patterns in the GA and NSGA-II algorithms, respectively, increased the storage of virtual water at an optimum of 29% and 35%. This model, as the decision support system, can play an effective role in deciding managers for different purposes. Also, the repeatability, runtime, and convergence of algorithms in the model indicate the superiority of the NSGA-II algorithm than GA. So that NSGA-II algorithm has less time in executing model, more convergence and less variance in the repeatability test than GA algorithm. This model can act as a decision support system to play an effective role in decision makers based on different goals. In this research, the use of meta- heuristic algorithms to optimize the allocation of users with the virtual water approach can be expressed in the thematic innovation of this research.

In recent years, the issue of improving the spatial resolution of thermal images in urban areas has been introduced as a new challenge. The purpose of this study is to use the impervious surfaces indices and vegetation indices to improve the spatial resolution of Landsat ETM thermal band over Tehran as a part of the study area. After the initial pre-processing, images obtained using the mean filter was simulated at spatial resolutions 120, 240, 480, 720 and 960 m. The relationships between these simulated imaged with the image simulated at the resolution of 960 m were calculated by the use of regression models.
These derived models, containing vegetation and impervious surface indices, were then used to simulation of surface temperatures in different pixel sizes. The accuracy of each output, has been evaluated using the thermal images of ETM and MODIS sensors.
The results showed that by increasing the spatial resolution, the errors increases while the gradient of error is not fixed. So that in all indices, there are more increasing in gradient of error when the pixel size goes to smaller than 240 meters.
Moreover, the best performance was obtained by combination of impervious surfaces indices and vegetation indices due to the enhancement of spatial resolution of thermal images in Tehran city.
Using the combination of these indices, the spatial resolution of the MODIS sensor can be reached to about 240 meters, while the absolute error value is less than 1 K Kelvin.

Urban growth and increased use of vehicles have led to an increase in air pollution, especially in large and industrialized cities in recent years. Because of the adverse effect of air pollution on human and other creatures, prediction and modeling of this complex phenomenon have the main concern of researchers during the last years. The purpose of this research is to design an air pollution prediction system to identify the contaminated areas in order to help the urban managers and planners to control and reduce the amount of contaminants. In the proposed system in order to predict the air pollution in different seasons, PCA-ANFIS model has been used. In this system, meteorological data and concentrations of pollutants are used to predict air pollution in Tehran over the next 24 hours. In addition, spatial parameters including height, topography and distance from the road are used to model the spatial distribution of air pollution. Comparing the results of PCA-ANFIS and ANFIS methods prove that the proposed model obtained higher accuracy in less processing time.

Cocurrent access to high spatial and temporal resolution imageries is essential in many studies. However, this will not be provided by using images from one sensor. To achive this goal, the incorporation of different satellites with high spatial (e.g., Landsat) and temporal (e.g., MODIS) images can be used. In present study, one of newest data fusion model, Enhanced spatial and temporal adaptive reflectance fusion model (ESTARFM) was evaluated with actual satellite data (OLI image). For emplementation and evaluation of this model, two different periods were selected (the first period selected between the days 204 to 220 and the second one were between the days 220 to 236). For evaluating the obtained results, OLI satellite images were used as a refrence data. Results show that ESTARFM not only improves the accuracy of predicted fine-resolution reflectance, especially for heterogeneous landscapes but it preserves spatial details also. The Coefficient of Determination (R2) of blue, green, red and near-infrared estimation bands with actual satellite data was 0.90, 0.91, 0.91 and 0.85 respectively, and the average Root-Mean-Square Error (RMSE) in four bands are 0.025, 0.030, 0.036 and 0.049 successively. In addition, a comparison between obtained NDVI from estimated reflectance values and observed NDVI, indicates outputs of ESTARFM have acceptable accuracy of (R2 =0.87 and RMSE =0.056). Thereby, this model can be successfully utilized to fusion images for enhancing the spatial and temporal resolution of reflectance.

This article develops and demonstrates a new quantitative modeling approach for environmental health impact assessment of traffic scenarios. For this purpose, two models based on hierarchical fuzzy inference system (HFIS) are developed. In order to develop HFIS for modeling the effect of transportation system on the PM2.5 concentrations, the data from an air dispersion model are utilized. There are several advantages to this approach such as modeling the spatial variation of PM2.5 with high resolution, suitable processing requirements, and consideration of interaction between emissions and meteorological processes. Moreover, the resulting fuzzy landuse regression (LUR) is capable of using accessible qualitative and uncertain data. In order to develop HFIS for modeling the impact of traffic-related PM2.5 on health, a metric derived from epidemiological studies is employed. The suggested model improved the metric capabilities by modeling the uncertainty of relationships among parameters and parameter value. Two solutions are used to improve the performance of both models. First, the topologies of HFISs are selected according to the problem. Second, used variables, membership functions and rule set is determined together through learning. We examine the capabilities of the proposed approach with assessing the impacts of three traffic scenarios to deal with air pollution in Isfahan, Iran and compare the accuracy of the results with representative models from existing literature. The models are first developed based on the current traffic conditions. Then; Low Emission-Zone and Odd/Even scenarios are examined. The examination shows that, they are the most and least effective scenarios in reducing air pollution and improving environmental health, respectively. The obtained results demonstrate that the proposed approach has desirable accuracy; beside that the model can provide better understanding of phenomena and investigating the impact of each of parameters for the planners.

Finding optimal Paths between two points on the Road network is one of the most spatial analysis in GIS. The high diversity of possible Paths between two points and difficult in apply all parameters simultaneously select the optimal Path (length of Path, easily track, traffic, road quality…) make finding optimal Paths problem to a difficult problem. Also, in some cases, two or more incompatible effective parameters such as length of the route and traffic adds to the complexity of the problem. Optimization algorithms, such as multi-objective genetic algorithm NSGA-II, that have ability simultaneous. Apply multiple incompatible parameters, can help GIS to solving these problems. Present a NSGA-II model on GIS based for finding optimal paths between origin and destination in the road network is the main Target of this paper. Also two GA innovative operator developed for enhance the ability of the model to find the optimal paths. Output of the model might be introduced optimal paths that they are shorter, quality of roads, transit of intersections and traffic. A hypothetical road network with the necessary restrictions, designed and utilizes for test the capabilities of the innovative model. Evaluation results show that the model is able to finding optimal Paths with multiple incompatible parameters.