Search

Satellite image analysis is a research area in which many research studies are carried out for civil and military applications in the field of image processing. Satellite imagery has many applications including recognition, detection and classification of regions, buildings, roads, aircraft and other man-made objects. Among these, especially aircraft detection is strategically important for military applications, and forms the basis of this study. In the first phase of the study, a new dataset of aircrafts is created from Google Earth images to compensate the shortage of data set in this area. In the second stage, the detection of air vehicles was carried out using algorithms based on Convolutional Neural Network (CNN). Region-based Fully Convolutional Network (R-FCN), Single Shot Multi Box Detector (SSD) and Faster R-CNN methods are used for this process. The obtained accuracy rate for R-FCN, SSD and Faster R-CNN are 98.01%, 69.71% and 96.56%, respectively., Uydu görüntü analizi, görüntü işleme alanında sivil ve askeri uygulamalar için birçok çalışmanın yapıldığı geniş bir araştırma alanıdır. Uydu görüntüleri, bölgelerin, binaların, yolların, uçakların ve diğer insan yapımı nesnelerin tanınması, algılanması ve sınıflandırılması olmak üzere birçok uygulamada kullanılmaktadır. Uçak tespiti özellikle askeri uygulamalar için stratejik öneme sahiptir ve bu çalışmanın temelini oluşturmaktadır. Bu çalışmada uçak tespiti uygulamalarına yönelik veri seti eksikliğini telafi etmek için Google Earth görüntülerinden yeni bir uçak veri seti oluşturulmuştur. Hava araçlarının tespiti için Konvolüsyonel Sinir Ağlarına (CNNs) dayalı algoritmalar kullanılmıştır. Uçak tespitine yönelik yapılan deneylerde Bölge Tabanlı Tam Evrişimli Ağ (R-FCN), Tek Atışlı Çoklu Kutu Detektörü (SSD) ve Daha Hızlı R-CNN (Faster R-CNN) yöntemlerinin performansı karşılaştırılmıştır ve sırasıyla % 98.01,% 69.71 ve % 96.56 doğruluk oranları elde edilmiştir.

Drone detection and classification, important in military and civilian applications, are performed using different sensor signals. Proposed study handles this task using Radio Frequency (RF) signals utilizing basic machine learning methods. It is composed of two main stages as feature extraction succeeded by training/testing of the model. In feature extraction stage, valuable information for classification, contained in the RF signal, is obtained. For this purpose, spectral features, frequently used in speech processing applications, are employed. Specifically, Power Spectral Density (PSD), Mel-Frequency Cepstral Coefficients (MFCC) and Linear Frequency Cepstral Coefficients (LFCC) are adopted by adjusting filter bank margins and parameters for this task. In the second stage, a Support Vector Machine (SVM) classifier is first trained based on the obtained features and finally tested to measure its performance. All experimental studies are carried out using publicly available DroneRF dataset. This dataset contains 2-Class, 4-Class and 10-Class samples for drone existence vs. background (BG), drone types and drone operation modes, respectively. The best classification results are obtained using, PSD, MFCC and LFCC based features for 2-Class, MFCC and LFCC based features for 4-Class and LFCC based features for 10-Class. Accuracy rates for 2-Class, 4-Class and 10-Class are 100%, 98.67% and 95.15%, respectively. These results show that the proposed method outperforms the results given in the literature for DroneRF dataset.

In this study, using the polysomnography data set, the classification of sleep stages was realized automatically with supervised learning method. In this study, sleep stage classification was carried out in three stages. In the first stage, the biomedical signal was divided into its independent components by Independent Component Analysis method. In the second stage, feature extraction was performed by using Mel Kepstrum Coefficient method. In the third stage, artificial neural networks were trained by using the extracted features and the sleep phases were estimated by using software architecture called Long ShortTerm Memory. As a result of the classification process performed in this way, the accuracy rate of the ten fold cross validation obtained for the binary classification (asleep / wake) was found to be 97,87%. For the five fold classification problem, the accuracy rate of the subject dependent algorithm and polysomnography data obtained using a single EEG channel was found to be 93,36% as a result of the ten-fold cross validation process.

Satellite image analysis is a field of study in the field of image processing for many civil and military applications. Satellite images have many civilian applications, including recognition, detection and classification of regions, buildings, roads, aircrafts and other man-made objects. Among these, especially aircraft detection is strategically important for military applications and forms the basis of this study. In the first phase of the study, a new dataset composed of airplanes at different airports on Google Earth's satellite images is prepared to eliminate the lack of data set in this field. In the second stage, the detection of aircraft was carried out by using algorithms based on Convolutional Neural Network. Transfer Learning was used to improve the performance of this process. The accuracy rate obtained as a result of Transfer Learning with previously trained network Alexnet, was 91%.

We investigate the effect of fabrication parameters (nonequal surface current densities, impurities inside the structure, etc.) and loss factor on reflectivity and transmittivity measurements from porous silicon Fabry–Perot cavities with finite-size substrate thicknesses. We apply the formalism based on wave cascade matrix method for obtaining dependencies of reflectivity and transmittivity. From our analysis, we note the following results. First, resonance behavior of reflectivity and transmittivity changes only when optical/physical properties of middle layers of the cavity alter. Second, for lossless cavities, while reflectivity and transmittivity considerably change with surface characteristics (optical/physical properties of first layers), transmittivity is significantly modified by a change of optical/physical properties of middle layers (microcavity region). Third, loss inside a FP cavity makes the transmittivity more immune to variations in optical/physical properties of middle layers. Finally and most importantly, transmittivity values at resonance wavelength as well as the resonance wavelength shift can be utilized for the identification of unknown chemical/biological molecules by lossless FP cavities. For validation of these results, we carried out reflectivity and transmittivity measurements from some arbitrarily chosen positions but around the center of two fabricated FP cavities resonating at 1456 nm and at 542 nm.

We propose a novel method for identification of various gases using combined measurements of wavelength shift ( $\Delta \lambda _{0}$ ) and reflectivity difference ( $\Delta R$ ). The method relies on plotting the measurement data on the $\Delta R$ – $\Delta \lambda _{0}$ space and exploits in the identification process the effect of significant loss behavior of bulk silicon near or within the visible range (for lossless gas vapors). It has been validated by reflectivity measurements on a Fabry–Perot cavity resonating around 745 nm for four different gas vapors whose reflective indices are close to each other and by two metrics for separability/identification analysis by the scattering matrix method. It is noted that the combined measurements of $\Delta \lambda _{0}$ and $\Delta R$ result in better identification than that obtained by the measurement of $\Delta \lambda _{0}$ (or $\Delta R$ ) itself.

Our research envisions a framework to analyze dynamics of service delivery over different environments including cloud platforms. Our research takes into account the business objectives, service development and deployment dynamics, data center and cloud dynamics for analysis, prediction, simulation and comparison of profits generated by the delivered services to the organization. This paper focuses on estimating the cost of digital services when such services are provided with on-premise IT infrastructure and with on-cloud infrastructure in a hybrid cloud deployment. The required computing resources for delivering the digital service are estimated. After that, the costs are estimated with pay-per-use pricing, subscription pricing and auction pricing models.

A free-space measurement method has been proposed for complex permittivity determination of lossy materials at millimeter and terahertz frequencies. The method relies on the Fabry-Perot interference effect between two identical samples separated by a varying air region. We derived a general metric function for fast and accurate complex permittivity determination of lossy materials using intensity data of transmittance power measurements. We have shown that via a graphical method, unique permittivity retrieval is possible using only two measurements of transmittance data corresponding to different distances between identical lossy samples. We have validated the proposed method by performing a numerical analysis.

In this study, an innovative design for porous silicon (PSi) based sensors is proposed to eliminate some problems of conventional PSi sensors such as undesired reflections from imprecise positioning of the fiber optic cable and the PSi structure. Our design has three stages as hole milling for fiber socket, fabrication of filter structure, and integration of the fiber optic cable to the bulk Si. We have carried out reflectivity measurements of alcohols with close refractive index values by the proposed and conventional sensors. From the measurements, we note that both sensors have equal sensitivity in identifying alcohols, that repeatability of our proposed sensor is relatively higher, and that our sensor allows easy positioning and flexibility in sensing applications. Nonetheless, our proposed sensor necessitates a thorough fabrication process and a methodological preparation.

In this paper, we investigate the effect of non-uniformities (enlargement of current passage, non-equal surface current densities, etc.) in axial as well as transverse directions of a porous silicon Fabry-Perot (FP) cavity as well as loss nature of bulk silicon on spectral properties of this cavity, even that cavity is created with an anisotropic etching process. Without correct and comprehensive characterization of such cavities by incorporating these non-uniformities and inherent lossy nature of a cavity, detection and identification of biological and chemical molecules by that cavity may yield unpredictable and misleading results. From our simulations, we note the following two key points. First, effects of the refractive index and the thickness of microcavity region of a lossless or lossy FP cavity on resonance wavelength is more prevailing than those of first and last layers. Second, the effect of some small loss inside the FP cavity is not detectable by the measurement of resonance wavelength whereas the same influence is noticeable by the measurement of reflectivity. We carried out some measurements from two different regions on the fabricated cavities to validate our simulation results. From a practical point of view in correct detection and/or identification of lossy biological or chemical vapor by FP cavities, we conclude that not only the measurement of resonance wavelength as well as its shift but also the reflectivity value at the resonance wavelength or some specific wavelengths should be utilized.

Three difierent techniques are applied for accurate constitutive parameters determination of isotropic split-ring resonator (SRR) and SRR with a cut wire (Composite) metamaterial (MM) slabs. The flrst two techniques use explicit analytical calibration- dependent and calibration-invariant expressions while the third technique is based on Lorentz and Drude dispersion models. We have tested these techniques from simulated scattering (S-) parameters of two classic SRR and Composite MM slabs with various level of losses and difierent calibration plane factors. From the comparison, we conclude that whereas the extracted complex permittivity of both slabs by the analytical techniques produces unphysical results at resonance regions, that by the dispersion model eliminates this shortcoming and retrieves physically accurate constitutive parameters over the whole analyzed frequency region. We argue that incorrect retrieval of complex permittivity by analytical methods comes from spatial dispersion efiects due to the discreteness of conducting elements within MM slabs which largely vary simulated S-parameters in the resonance regions where the slabs are highly spatially dispersive.