Gemilerin ısar ile görüntülenmesinde radar saçılma yüzeylerinin fizik optik yardımıyla modellenmesi
Başlık çevirisi mevcut değil.
- Tez No: 46368
- Danışmanlar: PROF.DR. ERCAN TOPUZ
- Tez Türü: Yüksek Lisans
- Konular: Elektrik ve Elektronik Mühendisliği, Electrical and Electronics Engineering
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 1995
- Dil: Türkçe
- Üniversite: İstanbul Teknik Üniversitesi
- Enstitü: Fen Bilimleri Enstitüsü
- Ana Bilim Dalı: Belirtilmemiş.
- Bilim Dalı: Belirtilmemiş.
- Sayfa Sayısı: 106
Özet
ÖZET Elektromagnetik teorinin son yıllarda oldukça yoğun biçimde uygulandığı alanlardan birisi de karmaşık hedeflerden saçılmanın modellenmesi ile tanınma ve sınıflandırma amaçlı SAR (Synthetic Aperture Radar) ve ISAR (Inverse Synthetic Aperture Radar) görüntüleme radarlarıdır. Cisimlerin yüksek çözünürlüklü radarlar ile görüntülenmesinde hedef yüzeyindeki baskın saçıcılar (saçıcı merkezler), hedefin saçılma karakteristiklerini verir. Bu karakteristikler yardımıyla saçılan işaretten karmaşık cisimlerin sınıflandırılması için gerekli olan (boy, yükseklik v.b.) nitelikler çıkarılabilir (feature extraction). Bu çalışmada, boyutları dalga boyuna göre çok büyük karmaşık cisimlerin radar saçılma yüzeylerin yardımıyla, sınıflandırmada kullanılacak karakteristiklerinin elde edilmesi amaçlanmıştır. Hedeflerin radar saçılma yüzeylerinin elde edilmesi için oldukça basit bir modelden yararlanılmıştır. Modellenmek istenen gemi benzeri hedefler büyük ölçüde dalga boyuna göre çok büyük eğrilik yarıçaplı yüzeyler içerdiğinden, hedef üzerinden saçılan alanın hesaplanmasında fizik optik (FO) yaklaşıklığı kullanılmıştır. Bilindiği gibi FO, hedefin aydınlatılan kısmında yüzey akımının gelen alan ile orantılı, aydınlatılmayan kısımda yüzey akımının sıfır olduğu temeline dayanır [l,sf 29]. FO yaklaşıklığı düzleme benzer yüzeyler için geçerli olduğundan hedef yüzeyinin modellenmesinde üçgen biçiminde düzlemsel elemanlar kullanılmıştır. Ayrıca modellenen yüzeyin kırık çizgilerle sınırlı olması durumunda, sınırın üçgen elemanlarla ifade edilmesinde (örneğin dörtgen elemana göre) daha az hata yapılmaktadır. Karmaşık hedefin modellenmesi ve buradan hareketle radar menzil profili ve ISAR modlarında elde edilecek hedef işaretinin belirlenmesi için“Görüntüleme Radaı Simülatörü”(GRS) geliştirilmiştir. Söz konusu algoritmanın işlem akışında hedef yüzeyi üçgen biçiminde düzlem yüzey elemanları ile temsil edilir. Her bir yüzey elemanından saçılan alan Bölüm 3'te verilen temel bağıntılar yardımıyla hesaplanarak hedefin toplam radar saçılma yüzeyi elde edilir. Çeşitli t anlarında hedef hareketine bağlı olarak elde edilen bu bilgiler uygun işaret işleme teknikleri kullanılarak hedefin ISAR (Inverse Synthetic Aperture Radar) görüntüsü elde edilir. Söz konusu algoritma EK-l'de verilmiştir.
Özet (Çeviri)
SUMMARY Numerous new advances have been made in electromagnetic theory in recent years. This due, in part, to new applications of the theory to many practical problems. For example, in microwave and milimeterwave applications, there is an increasing need to investigate the electromagnetic problems of new guiding structures, phased array antennas, microwave imaging, polimetric radars, microwave hazards, frequency-sensitive surfaces, composite materials, and microwave remote sensing. In radar the radio wave is transmitted toward an object and the scattered wave received by an antenna reveals the characteristics of the object, such as its position and motion. The identification and classification of complex objects whose dimensions are larger than the wavelength (of the incident wave) such as a ship utilizes radar cross section. On the other hand the main problem of identification and classification involves determining some basic features which forms basis for recognizing different objects. These features should be common for all targets and they also should be as simple as possible. The radar systems used for identification and classification purposes have high range resolution and/or cross range resolution. With such a modern radar system it is possible to obtain a range resolution and/or cross range resolution at the order of 1 meter. 2 So for large objects (e.g. ships) the scattered radar signal may be sampled at 10 - 10 resolution cells. Since each of these samples originate from a region whose position on the ships are known at least therotically a low resolution image of the object may be generated. The signal power coming from a resolution cell (i.e. the brightness of the pixels) strongly depends on many parameters. So its very difficult to find a similarity between the radar image and the optical image of an object. Most the time even a carefull and experienced operator can not realize a similarity between the radar image and its optical appeareance. In spite of the strong dependence of radar image to many parameters as mentioned above, some important characteristics of the target can be estimated with sufficient reliability due to additional data coming from large number of samples. These characteristics are the basic qualifiers which will be used to identify and classify the object such as the length, height, superstructure and position of the target. Knowing this qualifiers for a given target let one classify the object when some conditions are met. There are published algorithms for solve up the classification problem. But vuidentification and classification techniques are not investigated in order to keep the work bounded. Obtaining the radar image of a given object with an imaging radar system (ISAR) requires a certain level of knowledge and experience. Even worse, in order to use the obtained image data one needs a complete and reliable database. The recent development of computer technology in the last decade has made it possible to simulate such complex and costly measurement systems in totally artificial environment. Modeling and simulation from this point of view offers not only a method to change parameters simply by changing model data but also makes it possible to suppress second or higher order effects. The system can be simplified to a desired level of complexity. But one should keep in mind that every simplification in the model makes it less accurate. This work basically involves the evaulation of those characteristics that will be used in classification by utilizing radar cross sections of a complex object whose dimensions are much larger than the wavelength of the illuminating wave. In order to achieve such a model of the complex object the surface of the target is considered as a collection of surface elements. The main formulation in calculating the scattered field from a triangular planar surface element is briefly discussed below: When an object is illuminated by a wave (fig.l), a part of the inciden power is scattered out and another part is absorbed by the object. The characteristics of scattering phenomena, can be expressed more conveniently by assuming an incident plane wave. Let us consider a linearly polarized electromagnetic plane wave propagating in a medium with dielectric constant e0 and magnetic permability (i0 with the electric field given by; iff)**?-*** the amplitude of E, is chosen to be 1 (volt/m), and k is the wave number. X, is a wave length in the medium and T, is a unit vector in the direction of wave propagation. The total electric field in the distance R from a reference point in the object, in the direction of vector 5, consists of the incident field E{ and the field Es scattered by the particle. We can say for large distances; e-m R Milf@,T) is called the scattering amplitude. Keeping definition of radar cross section, o = lim 4ti £-=- in mind, calculating the bistatic radar cross section utilizing the scattering amplitude yields o = 4» U f For the monostatic scattering case one assumes; E,(T) Figure 1. Basic Scattering Mechanism PO approximation assumes the surface currents induced at the target surface are related to the in coming wave; J Jir) = )2(fixS) in the illuminated region 0 in the shadow If one inserts this surface current into the integral definition of the scattered field, obtains the bistatic scattered field : rx-m 2nR and using the relations given above the bistatic radar cross section expressing comes out; o(-U) jfc2 [ (A-t) e**1 dS Assuming the target surface S can be divided into N part which the phisical optics approximation is applicable and each part is independent of others. The total surface equals the sum of the surfaces all part S = £ AS, and the total radar cross section can be expressed as; A2 A Ti M / (A-İ) e*»' dSl AS, n,, in this expression denotes the unit normal vector of each surface component. The complex targets such as a ship mainly consists of large planar surfaces for which the PO approximation is applicable. On the other hand for surfaces which multiple reflections may take place the scattered field can not be calculated using the method explained above and should be handled seperately. The physical optics approximation is used in order to calculate the scattered field from the target simply because in most cases for large objects like a ship the radius of curvature of the target surface is much larger than the wavelength of the incoming wave. Since the PO approximation is applicable to planar surfaces, the surface element in modeling the target is chosen to be triangular. Another reason for this choise is the fact that the error introduced by triangular element when modelling surfaces with curved boudries is smaller in comparison to other surface element shapes (e.g., rectangular elements). In Chapter 3, test problems given below are considered: Rectangular flat plate (infinite radius of curvature in both directions) Cylindirical plate (infinite radius of curvature in one direction)* Spherical shell segment (finite radius of curvature in both directions) * Dihedral corner reflector (with multiple reflections) * Rectangular cylinder (as a simple three dimensional object) The results obtained from PO approximation are compared wiht the experimental results. In Chapter 3, the solutions obtained from moddeling the surfaces using triangular surface elements and the results are compared with the experimental results. In order to obtain the target image with a given resolution, the dimension corresponding to the radar pulse width should be much smaller than the target dimensions. So teh complex and large objects such as a ship should be divided into range cells. The radar image may be generated mainly by two methods; * Only range sampling * Both range and Doppler sampling The radars of the first kind are called profiling radar. Since the range resolution achived in this type of radars (or this mode of operation), typically an order of magnitude smaller than the dimensions of target, the scattering centers placed on the range direction at least a range cell apart can be detectedseperately. The ISAR image is obtained from both the Doppler and the range samples. This type of radars, stores the sequential signal ramming from the range cells. Than utilizing the fraquency domain methods, the cross range information is generated. This information comes from the relative radar to scatterer speed differences for different scatterers due to the ship's (pitch,roll,yaw) movements. ISAR mode generates a two dimensional or three dimensional image of the target while profiling mode only gives the one dimensional radar image. This mainly considers the 'Radar Imaging Simulator' which is developted to obtain simulated radar images of complex objects.“Imaging Radar Simulator”(IRS) algorithm may be summarized as follows; First the target surface is divided into triangular surface elements. The scattered field from each of the elements is calculated from the expression given below. After this step the total radar cross section of the target is calculated for different t (time) values by taking the movements of target in several directions into account. Then the resulting data array is processed by an FFT algorithm and the results are encoded by gray levels which finally gives the ISAR image of the target. The ISAR image of the object is given to the classification algorithm either directly or further processed utilizing appropriate signal processing techniques. The formulation used in modelling is discussed in Chapters 2 and 3. The detailed description of the algorithm used for the simulator is given in APP.l. XI
Benzer Tezler
- Menzil profili modunda çalışan radarla otomatik hedef sınıflama
automatic target classification with radars operating in range profilling
CENK GÖKBERK
Yüksek Lisans
Türkçe
1997
Elektrik ve Elektronik Mühendisliğiİstanbul Teknik ÜniversitesiElektronik ve Haberleşme Mühendisliği Ana Bilim Dalı
PROF. DR. ERCAN TOPUZ
- Gemi hareketlerinin incelenmesi ve gemilere uygulanması
Investigation of ship motions and applications to ships
TAHSİN TEZDOĞAN
Yüksek Lisans
Türkçe
2011
Gemi Mühendisliğiİstanbul Teknik ÜniversitesiGemi İnşaatı ve Gemi Makineleri Mühendisliği Ana Bilim Dalı
PROF. DR. METİN TAYLAN
- Geminin taşıdığı yük miktarına göre en uygun gemi hızının ve yakıt tüketiminin belirlenmesi: Ro-Ro gemileri üzerine bir uygulama
Determining optimal ship speed and fuel consumption depending on ship's cargo amount: An application on Ro-Ro ships
SELİM BAŞTÜRK
Yüksek Lisans
Türkçe
2017
DenizcilikKaradeniz Teknik ÜniversitesiDeniz Ulaştırma İşletme Mühendisliği Ana Bilim Dalı
YRD. DOÇ. DR. SERCAN EROL
- Gemilerin yapısal güvenilirliğinin birinci düzey güvenilirlik yöntemleri kullanılarak incelenmesi
Assesment of ship structural reliability with using first order reliability methods
ORHAN PEHLİVAN
Yüksek Lisans
Türkçe
2017
Gemi Mühendisliğiİstanbul Teknik ÜniversitesiGemi İnşaatı ve Gemi Makineleri Mühendisliği Ana Bilim Dalı
PROF. DR. AHMET ERGİN
- Gemi eğilme momenti ve kesme kuvveti kesit zorlarının bureau veritas kuralları ve nümerik yöntem ile analizi
Analyses of ship bending moment and shear force internal forces with bureau veritas rules and numeric method
ERHAN ASLANTAŞ
Yüksek Lisans
Türkçe
2013
Gemi MühendisliğiYıldız Teknik ÜniversitesiGemi İnşaatı ve Gemi Makineleri Mühendisliği Ana Bilim Dalı
PROF. DR. AYDOĞAN ÖZDAMAR