ارزیابی دقت مدل رقومی ارتفاع سنجنده پریسم/ ماهوارهای آلوس با مدل رقومی ارتفاعی Aster

یکی از اهداف مدل رقومی ارتفاع تولید نقشه‌های توپوگرافی است که نشان‌دهنده عوارض طبیعی اعم از رودخانه ها، دریاچه ها، کوه ها و عوارض مصنوعی مانند شهرها، جاده ها و پل ‌ها بوده و در مطالعات زیرساختی و استراتژیک از اهمیت بالایی برخوردار است. استفاده از تصاویر ماهواره‌ای یکی از راه های استخراج مدل‌های رقومی ارتفاع می‌باشد. هدف از این مطالعه بررسی نحوه استخراج مدل رقومی ارتفاعی تولید شده از تصاویر سنجنده پریسم همراه با فایل کمکی چندجمله ای منطقی موسوم به rpc می باشد. به‌منظور نیل به اهداف مطالعاتی از تصاویر ماهواره‌ای با بهره‌گیری از روش اکتشافی با تکنیک سنجش از دور استفاده شده است. تصاویر استریو ماهواره استر اخذ شده در تاریخ 2010/05/21 میلادی مورد استفاده قرار گرفته است، همچنین برای ارزیابی مدل رقومی ارتفاع از تصاویر استریو سنجنده پریسم که در تاریخ 2009/08/07 میلادی اخذ گردیده، استفاده شده است. نتایج نشان داد که rmse  به‌عنوان شاخص خطا برای مدل رقومی ارتفاع استخراج شده از prism، aster به‌ترتیب 3/66 و 6/8 متر می باشد. نتایج به‌دست آمده از انحراف معیار پیکسل ‌های تصاویر استریو سنجنده پریسم در جهت طولی1/9 متر و در جهت عرضی 2/3 متر و فاصله پیکسل ‌های مدل رقومی ارتفاعی 3 متر می‌‌باشد. که دقت مدل رقومی aster کمتر از اندازه پیکسل ‌ها یعنی کمتر از 15 متر می‌باشد، یعنی در جهت طولی 6 متر و در جهت عرضی 7 متر در پیکسل است که در مجموع 13 متر می باشد. نتایج خطای انحراف معیار منطبق بر نتایج rmse بود که تایید کننده مدل رقومی ارتفاعی prism است. بنابراین دقت مدل رقومی استخراج شده از تصاویر سنجنده prism بالاتر از aster می باشد. پیشنهاد می ‌شود در کل مرزهای کشور از مدل رقومی ارتفاعی با دقت بالا که با استفاده از مدل رقومی ارتفاعی تولید شده از تصاویر استریو سنجنده پریسم از ماهواره آلوس که همراه با فایل‌های ضرایب منطقی چندجمله ای (rpc) برای تصحیح هندسی تصاویر می‌باشد، استفاده شود.

Evaluate the accuracy of Digital Elevation Model (DEM)of PRISM / ALOS with ASTER

Extended AbstractIntroductionTopographic maps show natural and artificial features. natural features such as rivers, lakes, mountains, etc., Manmade features such as cities, roads and bridges. Using the satellite images is a way to extract digital elevation models. In general, there are two types of resolution in digital ground elevation models. üArea resolution: The dimensions of the length and width of each cell in the pixel grid is a digital elevation model that shows the minimum dimensions of the topographic features taken on the ground.ü Height resolution: represents the minimum elevation dimensions that the digital elevation model is able to display. For example, in the digital model of ground elevation with a resolution of 30 meters, elevation features less than 30 meters are not visible.The digital elevation model can be prepared for a region with different accuracy. The high accuracy of the digital elevation map provides more accurate estimates of the physiographic characteristics of the basin, but the preparation of such maps is very costly. PRISM sensor from ALOS satellite with three cameras: 1 Forward 2 Vertical 3 Forward, which is captured earth surface with the characteristics of the earth (low and high). Therefore, an object that is high above the ground is shown with other points on a flat surface. As a result, by imaging points from different angles, the elevation of those points can be obtained through adaptive mathematical calculations. The purpose of this study is to evaluate the accuracy of the digital elevation model generated by the PRISM sensor of ALOS satellite in comparison with the digital elevation model of ASTER and SRTM for Sarakhs border region (between Iran and Turkmenistan). MethodThe study area is located in northeastern Iran in the range of 35 to 38 degrees north latitude and 56 to 60 degrees east longitude and on the border between Iran and Turkmenistan in the border region of Sarakhs. The research method in this research has an exploratory aspect that the production and extraction of digital elevation model from PRISM sensor stereo images from Alves satellite and its evaluation is with digital model extracted from ASTER image. The digital SRTM model has a spatial resolution of 90meters, the digital ASTER model has a spatial resolution of 15 meters and the digital elevation model obtained from the PRISM sensor from the ALOS satellite is 5 meters. In this study, elevation control points using Google Earth and GPS have been examined. The algorithms used in this method to extract elevation information are the same as the algorithms used in the photogrammetric method. Elevation digital models are made from satellite images taken in pairs. The accuracy of digital elevation models of this method is perfectly proportional to the scale or resolution of satellite images.Results & DiscussionIn this study, we evaluated the digital elevation model from stereo satellite images of ALOS/PRISM satellite and compared it with the digital model of ASTER elevation and ground observations in the Sarakhs border region located on the border between Iran and Turkmenistan. In this study, the ability to generate a digital elevation model prepared from stereo images extracted from a PRISM sensor with a file of rational polynomial coefficients has been investigated, and we compared it with digital models extracted from stereo ASTER satellite and digital models extracted from SRTM. The results obtained from the digital elevation model are the accuracy of the digital elevation model produced by the pair of ASTER satellite images using a correlation between the two images of 0.47 pixels. Due to the spatial accuracy of the image pixels, which is about 15 meters, the accuracy of the digital model is less than the size of pixels, i.e. less than 15 meters, 6 meters horizontally and 7 meters vertically, which is a total of 13 meters. The results show that RMSE as error index for digital model of elevation extracted from ASTER and PRISM and ground observations are 7.46, 8.77, 3.66 and 6.8 meters, respectively. The results obtained from the stereo images of the PRISM sensor are the standard deviation of the pixels in the longitudinal direction of 1.9 meters and in the transverse direction of 2.3 meters and the distance between the pixels of the digital model is 3 meters high. Therefore, the accuracy of the digital model extracted from PRISM sensor images is higher than SRTM and ASTER. It is recommended to use a highprecision digital elevation model in all borders of the country, which uses a digital elevation model produced from stereo PRISM images from ALOS satellite, which is accompanied by polynomial logical coefficient (RPC) files for geometric correction of images. Conclusion The higher the accuracy of the DEM, the more efficient it will be and give border commanders the ability to make better decisions in different situations. The elevation accuracy obtained from the stereo images of the PRISM sensor is 3 meters. The accuracy of the digital model of SRTM elevation in the plains is about 30 meters, which can be used for studies of phase zero and one of the projects, as well as reducing the huge costs of studies. The results of this paper, shows that the accuracy of the digital elevation model produced from the stereo images of the PRISM sensor is higher than the digital elevation and SRTM digital models, i.e. the RMSE error and standard deviation are relatively lower. As a result, it is recommended for border studies that require higher accuracy, and the entire borders of the country, to use the digital elevation model with accuracy.