Please use this identifier to cite or link to this item:
|Title:||Körfezlerdeki su kalitesinin uydu görüntü verileri yardımıyla incelenmesi|
|Other Titles:||The Evaluation of water-quality in the bays by satellite images|
|Publisher:||Fen Bilimleri Enstitüsü|
Institute of Science and Technology
|Abstract:||Uzaktan algılama, yeryüzündeki kaynakların yönetimi, doğal ve kültürel çevrenin gözlenmesi ve haritalama amaçlarına yönelik olup, uzaya fırlatılan uydular aracılığıyla hızlı, doğruluklu ve konvansiyonel ölçme yöntemlerine kıyasla daha az maliyetli, detaylı ve/veya genel kapsamlı bir araştırma olanağı sunmaktadır. Bu çalış mada da, günümüzde özellikle yerleşim ve endüstri bölgelerinde tehlikeli boyutlara ulaşan su kirliliği ve buna bağlı olan kirlilik parametrelerinin incelenmesi esas alınmış olup İzmit Körfezi ile Finlandiya körfezinde seçilen bölgelerde su-kalitesi irdelenmiş tir. Çalışmanın ilk bölümünde, uzaktan algılamanın kısa bir tanımı verildikten sonra çalışmanın genel amacı ile değerlendirmede göz önüne alınan kriterler kısaca belirtilmiştir. Daha sonra elektromanyetik spektrum ve spektral özellikler (özellikle su cismi için) ile kısaca kontrollü ve kontrolsüz sınıflandırma yöntemleri açıklanmıştır. Veriler LANDSAT ve NOAA uydularından alındığından, bu uyduların özellikleri, taşıdıkları algılayıcılar, veri elde edilimi, algılama arızalan ve giderilme yöntemleri açıklanmaya çalışılmıştır. Dördüncü bölümde, kullanılan dijital veri değerlendirme sistemi, görüntü işleme yazılımı ve görüntü elde etme işlemi açıklanmıştır. Uygulama bölümü olan beşinci bölüm iki kısımda sunulmuştur. İlk kısım, LANDSAT-5 TM verileri ile İzmit körfezinde seçilen inceleme bölgesinde su kalitesi nin irdelenmesidir. Bu bölümde, bölgede TÜBİTAK ve İ.T.Ü. Çevre Mühendisliği Bölümü tarafından yapılmış olan su kalitesinin ve buna bağlı kirlilik parametrelerinin dağılımına ilişkin teknik raporlar, bölgedeki spektral verilerin dağılımını gösteren histogramlar, dijital görüntü zenginleştirme yöntemleri ile farklı bant kombinasyonla rı ve bant oranlarına uygulanmış olan sınıflandırma yöntemleri örnekler ve resimler ile sunulmuştur. Uygulama bölümünün ikinci kısmı, ardı sıralı 3 gün için elde edilmiş olan NOAA/AVHRR uydu verileri aracılığıyla Finlandiya körfezinde seçilen bölge için su kalitesinin değerlendirimini içermektedir. Bu bölümde, uydu verilerinin elde edildiği tarihler (25, 26 ve 27 Haziran 90) için bölgedeki gözleme istasyonlarından elde edilen yer doğruluklu ölçümler, uydu verilerinin spektral dağılımını gösteren histog ramlar, uydu verilerine kanal kalibrasyonunun uygulanması ve düşeye çevrilmesi, orijinal/kalibre edilmiş uydu görüntü verileri ile yer doğruluklu ölçümler arasındaki korelasyonu irdeleyen istatistiksel değerlendirimin lineer regresyon analizi yardımıyla gerçekleştirimi ile bölgenin sıcaklık ve klorofil-a parametreleri için yapılan sınıflandır ma sonuçlan grafik, tablo ve resimler aracılığıyla örnekler üzerinde açıklanmıştır. Son bölüm, çalışmada yapılmış olan iki uygulama örneğinin sonuç irdele mesini içermektedir. Bu çalışmanın uygulama bölümleri Helsinki Teknoloji Üniversitesi, Fo- togrametri ve Uzaktan Algılama Enstitüsü ile Finlandiya Ulusal Ölçme Birliği, Uzak tan Algılama Bölümünde gerçekleştirilmiştir. |
The Earth is the only planet known to support life. The Earth's energy ba lance, climate, biochemistry and cycling of chemical elements and water are affected by the abundance, distribution and function of terrestrial biota in processes of global significance will require interdisiplinary coorperation, a global perspective and new research strategies. Remote Sensing provides the global perspective and is destined to be a key tool in developing new strategies and methodologies. Remote Sensing is defined as acquisition of information about the conditi on and/or the state of target by a sensor that is not in direct physical contact with it. A Remote Sensing system consists of four compenents: a source, interactions with the Earth's surface, interaction with the atmosphere and a sensor. Electromagnetic radiation is the link between the compenents of the remote sensing system. Electro magnetic radiation occurs as a continuum of wavelengths and frequencies from short wavelength, high frequency cosmic waves to long wavelength, low frequency radio waves. In remote sensing, mostly, visible and near infrared radiation in the waveband 0,4 - 3 jam, infrared radiation in the waveband 3-14 jam and microwave radiation in the waveband 5 - 500 mm are used. The sun is a natural source of electromagnetic radiation. The radiant solar energy that reaches the Earth's surface is composed of direct solar beam and diffuse-sky radiation. The diffuse component results from propagation through the atmosphere. A portion of the total (direct + diffuse) inci dent solar energy that reaches the Earth's surface is absorbed by the surface and the rest is reflected back toward the atmosphere. The quantity and quality of the reflected component depends on the condition and state of the surface. The reflected radiation from the surface is measured as a function of wavelength by radiometers, scanners and spectrometers that are mounted on ground-based, airborne, or spaceborne plat forms. Detection of variations in reflectance between objects is dependent upon four interrelated factors; the radiometric resolution of the sensor, the amount of atmospheric scatter, the surface roughness of the objects and the spatial variability of reflectance within the scene. First, sensors vary in their ability to detect differences in radiance. For example, on board the satellite LANDSAT - 5 the scanning radiometer called the Thematic Mapper can detect 256 levels of radiance. Second, atmospheric scatter increases the amount of radiance received by the sensor for each object and as a result the contrast between objects is reduced. Third, surface roughness is of importance the surface needs to be rough enough to allow radiation to interact with the surface of the objects. Fourth, the spatial variability of the scene is of importance because the radiance recorded from an area of ground also contains radiance from the surroundings areas. The remotely sensed information recorded as a photographs or magnetic tapes or other data storage media, is then used to characterize the condition/state of the surface at the time of measurements. To archive this informati on in a digital image format is the most comman form of storage because of the increased amount of information that can be extracted from the data. A digital image xv format is a matrix whose row and column indices identify a point in the image and the corresponding matrix element value identifies the gray level at that point as shown in equation below where each element of the array is a discrete quantity: f(x,y)~ f(0,0) f(0,l) f(0,N-l) f(l,0) f(l,l) f(l,N-l) f(N-l,0)f(N-l,l) ".".'? f(N-l,N-l) To extact all the available information relevant to a given theme, the digi tal image processing techniques applied to digital image data. Digital image proces sing involves the manipulation and interpretation of digital images with the aid of a computer. This processing can be divided into the following four phases: - Preprocessing. - Image enhancement. - Image classification. - Data merging. The term "preprocessing" covers all those operations intended to correct known distortions in the image and to yield a document or support suitable for preliminary direct analysis. The energy levels recorded on board a platform are not a direct function of the terrain observed; they are modified by the atmosphere, sunlight and other sources of interference. To overcome these difficulties, all measurements must be corrected and all images rectified. In order to more effectively display or record data for subsequent visual in terpretation, image enhancement procedures are applied to image data. The objective is to increase the amount of information that can be visually interpreted from the data. The most commonly applied digital enhancement techniques can categorized as contrast manipulation, spatial feature manipulation or multi-image manipulation. 1. Contrast manipulation : Gray-level thresholding, level slicing, and contrast stretc hing. 2. Spatial feature manipulation : Spatial filtering, edge enhancement, and Fourier analysis. 3. Multi-image manipulation : Multispectral band ratioing and differencing, princi pal components, canonical components, vegetation components, and intensity - hue-saturation color space transformations. xvi The objective of image classification operations is to replace visual analysis of the image data with quantitative techniques for automating the identification of features in a scene. Image classification involves the analysis of multispectral image data and the application of statistically based decision rules for determining the land cover identity of each pixel in an image. There are two main approach for classificati on. First one, in the supervised approach, useful information categories are defined and then their spectral separability are examined; the other one, in the unsupervised approach, spectrally separable classes are determined and then their informational utility are defined. The classes due to based on the natural groupings in the image values. Therefore, the analyst must compare the classified data with some form of reference data (such as larger scale imagery or maps) to determine the identity and informational value of the spectral classes. For the supervised classification, there are three basic steps. In the training stage(l), representative training areas are identified and a numerical description of the spectral attributes of each land cover type of interest in the scene is developed. Hence, the overall objective of the training process is to assemble a set of statistics that describe the spectral response pattern for each land cover type. Next, in the classification stage(2), each pixel in the image data set is categorized into the land cover class it most closely resembles. If the pixel is unsuffi ciently similar to any training data set it is usually labeled "unknown". After entire data set has been categorized, the results are presented in the output stage(3). Three typical forms of output products are thematic maps, tables of full scene or subscene area statistics for the various land cover classes, and digital data files amenable to inclusion in a geographic information system (GIS). There are several classifiers for both type of classification process. Usually the most accurate supervised classifier which is also used in this study, is Maximum Likelihood classifier. în this classification process, first it is calculated the mean vector, variance and correlation for each land cover class in the training data, on the usually valid assumption that the data for each class are normally distributed. With this information the spread of pixels around each mean vector can be described using a probability function. Pixels from the whole data set are allocated to the class with which they have the highest probability of membership. Combining image data for a given geographic area with other geographi cally referenced data sets for the same area are called data merging. Other data sets might consist of image data generated on other dates by the same sensor or by other remote sensing systems. Since the launching of LANDSAT - 1 in 1972, remote sensing has proved valuable in various fields such as geology, hydrology, forestry, oceangraphy, etc. The current LANDSAT satellites named LANDSAT - 4 and - 5 (LANDSAT D' ) were launched into repetitive, circular, sun-synchronous, near-polar orbits. The satellite orbit is at an altitude of 705 km. Either the multispectral scanner (MSS) or the thematic mapper (TM) sensors are carried on this satellites. Direct transmission of MSS and TM data to ground receiving stations is made possible via the antennas onboard the satellite. The TM is a highly advanced multispectral scanner incorpora ting a number of spectral, radiometric, and geometric design improvements relative to the MSS. Spectral improvements include the acquisition of data in seven bands instead of four, with new bands in the visible (blue), mid-infrared, and thermal portions of the spectrum. Radiometrically, the TM performs its onboard analog-to digital signal conversion over a quantization range of 256 digital numbers. Geometri cally, TM data resolution is a 30 x 30 m instead of 79 x 79 m. The main advantages offered by the imagery collected from this LANDSAT satellites are ready availability for most of the world, repetitive multispectral coverage and minimal image distor tion. xvu LANDSAT data initially used to obtain a synoptic view of a large area of the Earth's surface for interpretation with aid of aerial photographic interpretation techniques. But today, due to remarkable advantage of the digital image processing of remote sensing data, the monitoring of marine environments especially coastal areas has become one of the most effective and uptodate observation techniques. Due to close proximity to urban and industrial areas, sea water is subject to increasing pollution. Besides LANDSAT satellite images, NOAA meteorological satellite images also have confirmed the ability to detect color variations in the water and in many cases to releate these to specific pollutants or sources for water-quality monitoring applications. This NOAA meteorological satellite has an orbit similar to that of the LANDSAT satellite which is near polar and sun-synchronous, at an altitude of 830 km. But these satellite carry an Advanced Very High Resolution Radiometer (AVHRR) which has a spatial resolution of 1,1 km and a swath width of 3000 km. Water quality can be defined in terms of the concentration of its chemical constituent relative to a variety of potential uses. The feasibility of using remotely sensed data for water quality measure ments must be assessed by considering the principles of light and water interaction. Spectral characteristics of water vary with wavelength and are the result of not only the molecular nature of water but also impurities within the water body. Variables that can affect remote sensing of physical water-quality characteristics are time of year, sun-elevation angle, aerosol and molecular content of atmosphere, specular reflection of skylight from water surface, roughness of water surface, film, foam, debris, or floating plants on water surface, water colour, water turbidity, reflectance and absorptance characteristics of suspended particles, multiple reflections and scat tering of solar energy in water, depth of water and reflectance of bottom sediments and submerged or emergent vegatation. By using with the appropriate sea-truth data, the satellite images are capable of providing a comprehensive, objective, rapidly acquired, wide area synoptic coverage of water bodies for water quality evaluations. In this study, water quality has been examined by using satellite images in the İzmit Bay (with a LANDSAT-5 TM image) and the Gulf of Finland (with 3 sequent NOAA/ AVHRR images) in the National Board of Survey, Remote Sensing Section, Finland. In the first part of this thesis, the spectral radiation characteristics (especi ally for water bodies) and spectral classification methods (supervised and unsupervi sed) have been explained. In the second part, a short explanation of the LANDSAT and NOAA satel lites programme, the characteristics of satellites and sensors, acquisition of data, and scanning distribances and its correction techniques have been given. In the following part, the digital image processing system and image pro cessing software, called DISIMP (Device-Independent Software for Image Proces sing), that provide comprehensive facilities for analyzing multispectral images from satellites have been described. The fourth part, application part of the thesis, has been divided into two parts. First part has contained the evaluation of water quality in İzmit Bay with LANDSAT - 5 TM data. For that, technical reports releated to water quality and the distribution of pollution sources in studied area, the examination of the histograms which show the distribution of the spectral data, digital image enhancement techniqu es (contrast manipulation, spatial feature manipulation and multi-image manipulati on), and classification methods (Euclidean Distance, Maxsimum Likelihood and xvin Linear Discriminant Analysis) applied to different band combinations and band ratios have been showed. Second part has involved the application done in the Gulf of Finland by using NOAA/AVHRR images. In this part, the distribution of the polluti on parameters and the sea-truth data collected in the same days when 3 sequent NOAA/AVHRR images (25, 26, and 27 June 1990) obtained, the examination of the histograms releated to the distribution of the spectral data, the calibration of the orijinal image data (for chlorophyll-a and temperature parameters), rectification of the images, the statistical evaluation of the orijinal/calibrated data with sea-truth data using linear regression analysis, and color-coded mapping of water quality parameters in studied area have been examined. The final evaluation of these two applications has been given in the last part of this thesis. Factors that must be evaluated in deciding upon satellite surveillan ce of water-quality are feassibility, cost and value of information content. Although satellite remote sensing may not compare favourably with conventional procedures, the results showed that satellite data can provide a reasonably good estimate for the evaluations of water supply which is a topic that must concern everyone on the surface of this planet.
|Description:||Tez (Doktora) -- İstanbul Teknik Üniversitesi, Fen Bilimleri Enstitüsü, 1991|
Thesis (Ph.D.) -- İstanbul Technical University, Institute of Science and Technology, 1991
|Appears in Collections:||Geomatik Mühendisliği Lisansüstü Programı - Doktora|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.