الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 176 |  |  |
| | | from | 176 | | |
Abstract
Synthetic Aperture Radar (SAR) has entered its golden age that with SAR being widely used in relevant and critical applications in our daily life. SAR is unique in its imaging capability in that it provides high-resolution two-dimensional images independent of daylight, cloud coverage, and weather conditions; it is predestined to monitor fixed and dynamic processes on the Earth’s surface in a reliable, continuous, and global manner. SAR systems have a side-looking imaging geometry and are based on a pulsed radar installed on a forward-moving platform. The radar system sends out high-powered electromagnetic pulses and sequentially receives backscattered signal echoes. Significant penetration can occur depending on the frequency band, so imaged objects and media must be modeled as volumes (e.g., vegetation, ice, and snow, dry soil). Full polarimetric Synthetic Aperture Radar (SAR) data offers plentiful information about geology, geomorphology, hydrology, land cover, and soil classification and therefore helps in assessing their potential for development. In this study Radar signal has been processed to extract valuable information for various applications in our life, so Full - polarimetric ALOS/PALSAR-2 images as a Single Look Complex (SLC) product format with different dates were used and validated with numerous methods, that two case studies have been taken as follows; First; full-polarimetric SAR (ALOS/PALSAR L-band sensor) data is used for mapping different geological units and land features, which are covering some parts of the Greater Cairo area, Egypt. This study includes data collection, data interpretation, and a validation tool, as well as methodology and preliminary findings. The ALOS/PALSAR images were extracted, decomposed, filtered, and geo-referenced. An unsupervised classification scheme with 5 classes was performed for Radar data using (Wishart H-A-α unsupervised classifier), and it was exposed to the supervised classification technique with the assistance of the published geologic and geomorphological maps and the high-resolution Landsat images ratio techniques. These five classes are categorized as follows; Urban areas: which cover 22% of the study area, Agriculture: canopy covers 53% of the study area, The Nile: about 73 km of the Nile passing through the study area, and two geological Units, Te (Tertiary, which represents sands and sandstones with clay and marl), and Tp (Tertiary Pliocene, which represents sands, sandstone, and gravels). Moreover, the classification accuracy assessment (CAA) was performed for the obtained results using 327 ground control points. The CAA showed classification accuracy of around 81.82% with a Kappa coefficient of 0.8344. Moreover, a new model is proposed for classifying PolSAR images applied to the previous data which represent a part of the Greater Cairo area, Egypt. Temporal convolutional networks (TCN) deep learning has been used to extract the features from coherency and covariance elements and then train the model. At last, the SoftMax classifier was used to classify the PolSAR image. The proposed model was tested with evaluation metrics. The obtained results show that the proposed model can achieve high classification accuracies of 94.90%, 93.60%, 93.50%, and 93.10% for the extracted clases; Tertiary, Agriculture, Urban Area, and Water, respectively. with overall accuracy of 93.84%. Second; full polarimetric synthetic aperture radar images have been used in imaging the internal anatomy and dynamics of coastal dunes, which are very crucial to protect the environment. In this study, the optical and synthetic aperture radar (SAR) images were processed and integrated to extract information about the past, current, and future behavior of the coastal dunes of the northwestern part of the Nile Delta. The full-polarimetric ALOS/PASAR-2 images of years 2015 and 2017 were processed to generate the coherence change detection (CCD) and Pauli RGBs maps. The CCD images show very low coherence for the sand dunes, which means these dunes are dynamic. Therefore, the two generated Pauli RGBs maps subtracted to show the amount of changes in the scattering mechanism of the coastal dunes. The dunes show Bragg scattering because of the radar signals have been totally attenuated within the dry sand. Some areas have shown clear enhancement in the volume scattering related to increase in the vegetation cover. Moreover, the changes in both power and phase imbalances of different dune fields during the investigated years (2015 & 2017) show high differences in values, which reached 40 dB and 0.06 degrees, respectively. This might be related to changes in these dunes and not only because of the noise of the SAR system. In addition, the conventional change detection method using the supervised Landsat-8 images of years 2015 and 2017 were generated and show fairly change in the sand dune cover. Therefore, the high-resolution images of Google Earth were used to digitize the dune crests and measure their encroachment rate, which has reached about 4 m/year with the NW–SE direction. Such southward dunes encroachment offers room for the Mediterranean Sea to transgress and merge with the shoreline. The extracted radar and field studies show two successive groups of coastal sand dunes sexist in the study area; the old fixed dunes with high content of heavy minerals and difficult to be blown by the wind and the recent windblown longitudinal and barchan dunes with very low heavy minerals content. The old fixed coastal dunes will protect the properties from sea level rise and wave storms. While the movable recent longitudinal and barchan dunes are threating the surrounding projects and should be considered for mitigation actions. Finally, this research shows that by using full-polarimetric SAR data, the land cover and geology can be accurately mapped without suffering and wasting time, effort, or facing hazards, also processing full polarimetric SAR data produce a new applied change detection method for studying sand dunes worldwide for its clarity, simplicity, and relatively high resolution.