Monitoring of Changes in Salt Lake Maharlu Usinf Multi-Spectral Satellite Image Processing Techniques

Document Type : Research Paper


1 Associate Professor of Geography and Urban Planning, University of Mohaghegh Ardabili

2 Graduate student in Geomorphology University of Mohaghegh Ardabil

3 Professor in Remote Sensing and GIS of Tehran University


Monitoring of salt lake dry areas is an important parameter for sustainable development and environmental protection. Monitoring of dried areas of the salt lake requires extraction of thematic maps at different times. Powerful tool for remote sensing of the earth different ecosystems, such as Playa environments, is to produce valuable and useful data. The major causes of widespread applications of remote sensing data are simple, fast and useful and valuable for a variety of indoor environments. Despite the available images, their use in geomorphological phenomena is very limited in a variety of fields such as playa salt lakes in arid and semiarid areas. Limitations on the use of satellite data to map the areas affected by the salts depends on the spatial distribution of salts on the surface, the changes in salinity, vegetation as a barrier and spectral mixing with other levels of the ground. Some compounds are mixed with saline playa lakes and separate the salt basin using satellite images processing. In low humidity conditions, the amount of land affected by salt range reflects more visible light and low reflectivity, and, for example, water is represented in the mid-infrared bands. In order to reduce the adverse effects on salt and increasing the information about them, at least two bands of the sensor can be combined to create the composition and formulation.  This is the simplest way to interpret the results of visual computing the thresholds. The boundary thresholds below which the spectral space of a few pixels is supposed  are assigned to unknown or non-categorized classes. The main purpose of this study was to assess common methods and provide a novel method for extracting the salt lakes areas by the satellite images to monitor the occurrence of spatial extent of Playa. Given the importance of salt playa lakes in this area of research, monitoring of salt Lake Maharlu was performed using the ASTER images and TERRA satellite in 2010.
Materials and methods
The ASTER image data used in this study was taken in 2010. The sensor has wide spectral region with 14 bands from the visible to the thermal infrared in high spatial and spectral resolution. According to a study that evaluates conventional separation methods for Salt Lake Maharlu and thematic mapping using satellite image processing, the proposed method in this research include visual interpretation of false color composition and also the ratio between the band and threshold on the histogram of the image. Due to the complex interaction of a range of other phenomena in Salt Lake Playa Maharlu, it is required to employ output with high precision and to evaluate the performance of each of the above methods. The primary method is generating False Color Composite (FCC) and the Optimal Combination of Bands (OIF) through visual interpretation. By combining different bands of the ASTER application ILWIS, a large number of false-color images are used that the best color combination for the separation is combination of the bands 4,3,2(243RGB).. After producing the best color combination, the different detection techniques developed by linear histogram and edge detection filter, image contrast, and visual interpretation determined salt lake boundaries using the basic elements of visual interpretation (such as texture, tone, shape, color, function, shadows, location, etc.). The second way to distinguish the threshold on Runway 5 Salt Lake takes advantage of other phenomena.. Given that most of the other bands in the visible range are affected by the surface properties of salt, pure salt do not seem very good to separate the pixels. But mid-infrared bands of the area are more sensitive than salt water and salts are absorbed by water molecules rapidly than the visible bands are influenced by the surface properties of salt in the dry areas. so the bands of pure salt isolation of pixels from other pixels are better. Thus the separation of pure salts pixels from other pixels, the area of salt as they are not as Playa non salt, so the ASTER band 5 was chosen as the runway for extracting salt lake. In the third method of operation rating the ASTER spectral bands were used to extract the salt lake. The spectral characteristics of salt lakes with mean brightness values of image pixels can be mapped spectral reflection curves of the salt lakes and the formulas obtained and the desired parameters can be extracted. . Most bands that define the parameters used are salt lakes, in the range of visible and infrared bands are intermediate. The reason salt lake reflects the strong absorption in the visible and mid-infrared portion of the electromagnetic spectrum is so severe they had to be able to provide information parameters to increase the salt lakes. In this study, two of the indicators, new indicators of lake salt ratio (RSCI) and indicators of lake salt differential normalized (NDSCI) the characteristics of the ASTER dry and arid desert to distinguish this phenomenon from satellite images are presented, each characteristic, weaknesses and special abilities of their own. To evaluate the resulting maps, as a ground truth map for the semi-automatic method for measuring the accuracy of the map was produced, thus creating a ground truth image, the overall accuracy of the map was generated calculated. This accurately reflects accurately defined threshold for regional-scale studies of salt.
Discussion and conclusions
The results showed that, depending on the satellite image of the imaging phenomena, the spatial resolution is less possible in mixed pixels increases, Therefore, the high spatial resolution of the image can be more accurately position the salt lakes in arid regions can be extracted reliably and accurately. Method of visual interpretation of satellite images, especially color composite images using different bands will be a quick and comprehensive view, In this way an accurate depiction of salt lakes, especially in the border areas with varying combinations of vegetation and soil types are difficult and complex And identify the precise location of the boundary line of salt lakes are typically associated with errors. One of the fundamental problems with this approach, the boundary pixels are correctly detected This can be a major cause of medium or low spatial resolution sensors are used, as this would lead to mixed pixels. Threshold on the histogram using the methods that can be extracted from satellite imagery of the salt lakes. This is a reflection of the salt compared to other phenomena in the mid-infrared bands, and very small allotment is close to zero; therefore to extract the salt lakes, on the threshold of the action takes place in the middle infrared bands. Although this method has high accuracy, but it can be easily extracted automatically and quickly raised. Use the bands to extract the salt lakes is also difficult because of the different coatings than in places where the ground does not have an acceptable result. Due to the complex nature of this study provide a new way to Playa Maharlu and tested. This method is based on the combination of the two thresholds and ratio between the bands. The results clearly show that by using simple visual computing easily be extracted useful information from satellite images.
This study detected Salt Lake Playa Maharlu with the ASTER 2010images were processed, It was found that both the band and threshold rating range of the resulting image histogram for monitoring and mapping of salt lakes in arid areas is practiced. The results also showed that the overall accuracy indices NDSCI & RSCI and respectively 0.87 and 0.92 in saline lakes in arid and monitor resolution satellite imagery are effective.


Main Subjects