عنوان مقاله [English]
Radar interference technique is an efficient method in measuring ground surface displacement. So that with the use of this technology, it is possible to monitor small movements of the earth's surface continuously, with high accuracy and in a wide range. This technology has become very common in the study of natural disasters of the earth, including slope displacement, subsidence, earthquakes and volcanic activity. This technique compares the phase taken from two radar datasets at two different times and, by creating an interrogram, is able to measure changes in the earth's surface over time. In this study, in order to identify and measure landslides, radar images of 2015 and 2021 were used. In order to process the information, SARSCAPE software has been used, which has an estimated maximum landslide of 34 cm. Land use map of the study area was extracted using Landsat 8 image and using object-oriented classification. The results of this study showed that radar images have a good potential for detecting the instability of slopes and calculating their displacement. The results of adapting the land use map to the landslide map showed that the highest amount of slope movements are in the vegetation uses with 33 cm and the range with 28 cm, respectively, with the lowest value to the residential area with 9 cm. Sudden landslides and the destruction of vulnerable structures are possible accidents resulting from the effects of sloping movements that cause human casualties in urban areas. In some cases, these accidents can cause heavy and irreparable losses due to high population density or the expansion of the radius of the collapsed lands.
Continuous monitoring of land surface changes and identification of areas prone to slip movements, especially in the area of human settlements and communication infrastructure such as roads and railways, is one of the most effective factors in reducing casualties and natural hazards such as landslides and slopes. So far, several techniques have been proposed such as using the Global Positioning System, geodesy and tachometry, mapping cameras, laser scanning and lidar to monitor surface changes. However, due to the high cost of implementation, time consuming and limited coverage of the use of these methods, in the limited, the use of these methods in a wide range is not cost effective. But in addition to these methods, the radar interference technique with the ability to work in all weather conditions and the duration of day and night and with the ability to cover the ground and high spatial and temporal resolution, today is one of the most accurate (in millimeters) and least expensive Remote sensing techniques for detecting and monitoring surface changes, slow and unstable movements of amplitude around the world.
In this study, Sentinel 1 images were used to capture images in the C-band range of microwaves. Then the necessary processing was performed through SARSCAPE 5.2 plugin in ENVI 5.3 software and the technique used in this research to determine the amount of amplitude movements is a differential interferometric method with a combined opening of two frequent or non-frequent passes. One of the most basic steps in radar interferometry processing is to select the right image pair. Several factors such as sensor frequency, spatial baseline, temporal baseline as well as spatial overlap in the direction of sensor movement are effective in selecting image pairs. In the present study, two Sentinel-1-A images of SLC type related to 2015 and 2021 were used.
In this research, in order to extract the land use map, Landsat 8 satellite images of 2021 and June were used. ENVI 5.3 software was used for atmospheric and radiometric corrections and ARCGIS 10.5 software was used to extract the relevant maps. Object-oriented classification method was used in eCognition Developer64 software to land use classification. To estimate the classification accuracy, sample points taken from Google Earth images were used. In the object-oriented classification method, spectral information is merged with spatial information, and the pixels are segmented based on the shape, texture, and gray tone of the image at a specific scale, and the image is classified based on these components. In pixel segmentation, pixels are segmented by different algorithms, with different spectral and shape ratios and based on spectral and spatial properties in the form of various objects. During this process, image objects appropriate to homogeneity or heterogeneity were created based on the parameters of scale, color, shape, softness and compaction shape (Feizizadeh and Hilali 2010). Trial and error determined the best size that represents objects with different dimensions. In this study, a fragmented image with a scale of 45, Figure 0.5, compression coefficient of 0.6 was performed. After specifying the number of required classes in the classification, the training samples were determined and applied to the image surface in the software. Then the classification was done using the nearest neighbor and the decision tree method. It is necessary to use any kind of subject information, to know its accuracy and correctness. Classification accuracy in 2021 with 97% overall accuracy and 95% kappa coefficient.
The results of this study showed that radar images have a good potential for detecting the instability of slopes and calculating their displacement. The maximum amount of material movement is in the range of 34 cm. Which indicates that the area is active in terms of amplitude movements. Land use maps using Landsat 8 image were used using object-oriented classification in the study area. The results of matching the land use map and landslide map in Table 3 showed that the highest amount of landslide in the highest amount of amplitude movements, respectively, in vegetation uses with a value of 33 cm and rangeland with a value of 28 cm, respectively. They are located in a residential area of 9 cm. Sudden landslides and the destruction of vulnerable structures are possible accidents resulting from the effects of sloping movements that cause human casualties in urban areas. In some cases, these accidents can cause heavy and irreparable losses due to high population density or the expansion of the radius of the collapsed lands.