Uzaktan algılama verileriyle orman yangını analizi

Abstract

Uzaktan algılama biliminde, uydular aracılığıyla yer yüzündeki cisimlerin yayınladığı ve yansıttığı elektomanyetik enerji dijital olarak kayıt edilir. Kaydedilen bu enerji bilgisayar ortamında işlenip, analiz edilerek yeryüzündeki kaynakların yönetimi, doğal ve kültürel çevrenin gözlenmesi olanaklıdır. Uzaktan algılamanın bu özelliği; hızlı, doğru ve konvansiyonel ölçme yöntemlerine göre daha ekonomik ve ayrıntılı bir araştırma olanağı sağlar. Bu çalışmada ülkelerin en büyük doğal zenginlik kaynağı olan ormanların başlıca düşmanı olan yangın, yangın sonrası oluşan hasarın belirlenmesi amacıyla uzaktan algılama teknikleriyle incelenmiştir. Bu çalışmada ilk olarak uzaktan algılamanın kısa bir tamım yapıldıktan sonra çalışmanın genel amacı ile değerlendirmede göz önüne alman kriterler ve elektromanyatik spektrum, atmosfer etkisi, cisimlerin spektral özellikleri açıklanmıştır. Uydu sistemleri bölümünde, incelenen bölgeye ait dijital görüntülerin elde edildiği SPOT, LANDS AT ve İRS İC uydularının ve taşıdıkları algılama sistemlerinin özellikleri açıklanmış ve bu özellikler kapsamında sistemler birbiriyle karşılaştırılmıştır. Daha sonra dördüncü bölümde, dijital görüntü, dijital görüntü işleme teknikleri ve bu teknikler kapsamında görüntü zenginleştirme ve sınıflandırma işlemleri anlatılmıştır. Coğrafî Bilgi Sistemlerinin verildiği beşinci bölümde, CBS'nin tanımı, bişenleri ve bu çalışmada kullanılan CBS yazılımı ARCVDEW hakkında bilgi verilmiştir. Uygulama bölümünde, bölgenin coğrafi tamını, iklim ve bitki örtüsü ve jeolojik yapısı verilmiş yapılan diğer işlemler sırasıyla anlatılmıştır. Öncelikle mevcut 1/2S000 ölçekli topoğrafik paftalarla yangın sonrası durumu yansıtan tematik orman haritaların sayısallaştırılması anlatılmıştır. Daha sonra oran görüntülerle yapılan çalışma anlatılmıştır. Yangın öncesi ve sonrası yanan bölgedeki yansıtım farklılığım ortaya koymak için yapılan spektral profil analizi şekilleriyle anlatılmıştır. Sınıflandırma çalışması, uygun bant kombinasyonlarının seçimi ve görüntülerin sınıflandırılmasıyla anlatılmıştır. Bu işlemler sonucunda elde edilen veriler (oran, sınıflandırma ve orijinal görüntü verileri) oluşturulacak CBS ortamında birleştirilmek üzere ülke koordinat sistemine (UTM) dönüştürülmüştür. Son bölümde yapılan çalışmanın sonuç irdelemesi yapılmıştır.

The earth is the only planet known to support life. The earth's energy balance, climate, biochemistry and cycling of chemical elements and water are effected by the abundance, distribution and function of terrestrial biodata in processes of global significance will require interdisciplinary cooperation, 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. Major problems are geographic in nature; they involve vast areas of land and sea, climatic systems, regional distribution of water, vegetation, soil and rocks, lines of physical communication and trade, political, ethnic and language boundaries. The bulk of knowledge about the problems, their causes and possible solutions comes mainly from investigations of tiny and often isolated examples. A change for the better demands that the issues are addressed on the scale at which they really present themselves. Remote sensing does not directly hold forth solutions but supplies a publicly accessible pool of information that matches the interconnectedness, diversity and scale of global problems in a way that has never been possible before the last two decades of the twentieth century. It maximizes diversity of information and area of coverage at minimal cost and allows the interpreter to address several problems far more rapidly than by any other method. Remote Sensing is the collection of information about an object without coming into physical contact with it. A remote sensing system consists of four components; a source, interaction with the earth's surface, interaction with the atmosphere and a sensor. The principal advantages of remote sensing are the speed at which data can be acquired from large areas of the earth's surface and the related fact that comparatively inaccessible areas may be investigated in this way. The uses, both existing and potential, of such data within the various environmental disciplines are legion. The applications are generally as fallow;.? Meteorology; e.g. profiling of atmospheric temperature, pressure and water vapor content, measurement of wind velocity..- Oceanography; e.g. measurements of the sea surface temperature, mapping ocean currents and wave energy spectra..? Glaciology; e.g. mapping the distribution and motion of ice sheets and sea ice, determining the navigability of sea ice. - Geology, geomorphology and geodesy; e.g. identification of rock type, location of geological faults and anomalies, measuring the figure of the earth and observing tectonic motion... Topography and cartography; e.g. obtaining accurate elevation data and referring them to a given coordinate system, production and revision of maps. .. Agriculture, forestry and botany; e.g. monitoring the extent and type of vegetation cover and its state of health, identifying the host plants of pests, mapping soil type and determining its water content, forecasting crop yields... Hydrology; e.g. assessing water resources, forecasting meltwater run off from snow..? Disaster control; e.g. warning of sand and dust storms, avalanches, landslides, flooding etc., monitoring the extent of floodwater, monitoring of pollution. - Planning applications; e.g. generation of inventories of land use and monitoring changes, assessing resources, performing traffic surveys..- Military applications; e.g. monitoring the movement of vehicles and military formations, assessing terrain. Remote sensing of the environment comprises the measurement and recording of electromagnetic energy reflected from or emitted by the Earth's surface and atmosphere from a vantagepoint above the Earth's surface. Sensors mounted on aircraft or satellite platforms measure the amounts of energy reflected from or emitted by the earth's surface. Electromagnetic radiation is the link between the components of the remote sensing system. Electromagnetic radiation occurs as a continuum of a wavelength and frequencies from short waves. In remote sensing, mostly, visible and near infrared radiation in the waveband 0.4-3 urn, infrared radiation in the waveband 3-14um and microwave radiation in the waveband 5-500 mm is used. The Sun is a natural source of electromagnetic radiation. Electromagnetic energy reaching the Earth's surface may be reflected, transmitted or absorbed. Reflected energy travels upwards through, and intersects with, the atmosphere; that part of it which enters the filled of view of the sensor is dedected by the sensor and converted into a numerical value to be transmitted to a ground receiving station on the Earth. The amount and spectral distribution of the reflected energy is used in remote sensing to infer the nature of the reflected energy is used in remote sensing to infer the nature of the reflecting surface. A basic assumption is that individual targets (soils of different types, water, with varying degrees of impurities, rock of differing litologies, and vegetation and of various species) have an individuals and characteristic manner of interacting with incident radiation which is described by spectral response of that target. Detection of variations in reflectance between objects is dependent upon for interrelated factors. The radiometric resolution of the sensor, the amounts of atmospheric scatter the surface roughness of the objects and the special variability of reflectance within the scene. The remotely sensed information recorded recorded as photographs or magnetic tapes or other data storage media, is used to characterize the condition/state of the surface at the time of measurements. To archive this information in a digital image format is the most common form of storage because of the increased amount of information that can be extracted from the data. Image processing is generally considered to consist of the three steps of preprocessing, image enhancement, and classification. It is used to extract all the available information relevant to a given theme. Digital image processing techniques are applied to digital image data. A digital image 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 each element of the array is discrete quantity: «Xy) = W>) f(0,l) f(0,n-l) *1,0) f(l,l) f(lsn-l) fin-l,0)f(n-l,l) fi>l,n-l) Each region of the image is called a picture element, or pixel. The purpose of preprocessing is to remove systematic errors from the data. The most important preprocessing operations are the correction of radiometric and geometric errors, i.e. calibration of the detected signal and registration of the image data with true surface positions. The goal of Image enhancement is to increase the amount of information that can be visually interpreted from the data. The operations of image enhancement are contrast manipulation (i.e. gray level thresholding, level slicing, contrast stretching), spatial feature manipulation (i.e. spatial filtering, edge enhancement, Fourier analysis), multi image manipulation (i.e. multispectral band rationing and difference, principal components, vegetation components, intensity-hue-saturation color space transformations). Image classification, is the process whereby an image is converted into some kind of thematic map, in which regions with similar properties are indicated in the same way. It is a problem of recognition in that the numerical values associated each pixel are normally required to be identified in terms of an observable earth surface cover type. There are two main approaches for classification: Supervised and unsupervised. In supervised classification useful information categories are defined and then then- spectral separability are examined. In unsupervised classification spectrally separable classes are determined and then their informational utility are defined. For the supervised classification, there are three basic steps. In the training step 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. In classification step each pixel in the image data set is categorized into the land cover class it most closely resembles. After entire data set has been categorized the results are presented in the output step. There are several classifiers for both type of classification process. Most accurate supervised classifier is Maximum Likelihood classifier. In this classification process, the mean vector, variance and correlation for each land cover class in the training data are firstly calculated. By this it is assumed that tile 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. Remote sensing systems may be classified in a number of ways. The most useful distinctions we can draw are between active and passive systems and between imaging XI and do not imaging systems. We may also distinguish sensing by the wavelength of the radiation to which they respond. Active systems illuminator the object of study with their own supplied radiation whereas passive systems sense naturally occurring (emitted thermal or reflected solar) radiation. An imaging system slightly harder to define. It is a system, either active or passive, to which measures the intensity of the radiation reaching and which does so as a function of position on the earth surface so that a two dimensional pictorial representation of the intensity can be constructed. A non-imaging system is one, which either does not measure radiation intensity or does not do so as a function of position on the earth surface. After the first Landsat had been launched, remote sensing has proved valuable in various fields such as geology, hydrology, forestry etc. The current Landsat satellites named Landsat 4 and 5 were lunched into repetitive circular sun synchronous near polar orbits. The satellite orbit is at an altitude of 705 km. Either the multispectral scanner or thematic mapprer are carried on the satellites. Direct transmission of MSS TM data to ground receiving stations is made possible via the antennas on board the satellites. The TM is a highly advanced multi spectral scanner incorporating a number of spectral radiometric and geometric design improvements relative to the MSS. Spectral improvements include data acquisition of data in seven bands instead of four, with new in the visible (blue), mid infrared, thermal portions of the spectrum. Radiometrically, the performs its on board analog to digital signal conversion over a quantization range of 256 digital numbers. Geometrically, TM data resolution is a 30x30 m instead of 79x79 m. Indian Remote Sensing (IRS 1C) was lunched in April 1995. And it has been designed as a global mission to ensure availability of data all over the world. The IRS 1C satellite is placed in a polar sun synchronous orbit of 817 km. The IRS 1C has a unique combination of payloads consisting of three cameras; one operating in the panchromatic and the other to in the multi spectral bands these cameras are panchromatic camera (a resolution of 5.8 m), linear imaging self scanner (LISS 3) white filled sensors (WIFS). Using remotely sensed data to evaluate natural forest resources is one of its most challenging applications. The principle reason for this difficulty is the lack of homogeneity in the surface cover in term of tree type density, height. Many different species may be represented and where their crown size is less then the resolution dimensions of the system employed each pixel represent mixture. One of the greatest hazards in forested or scrublend areas is fire. For the forested areas the blue and red light absorbed by chlorophyll and other pigments. There is a slight reflectance peak in the green, which is the reason that growing vegetation appears green. Reflectivity rises sharply in the region of 0.75 urn and reflection remains high in the near infrared region between 0.75 and 1.35 um. This high reflection coefficient is due to the internal leaf structure. Between 1.35 and 2.5 um internal leaf structure has some effect but the refflectans is largely controlled by leaftissue water content. Spectrally, a soil or rock is very much different from a plant. Their spectral reflections rise steadily with wavelength a peak and then false of once more without having a sudden jump at the red infrared boundary. xu In this study, the forest fire in Marmaris, province of Mugla has been examined by using satellite data. The images used are acquired from Landsat, Spot and IRS-1C satellites. In the first part of this thesis electromagnetic energy and the atmospheric effects to this energy, the spectral characteristics of the objects (especially for vegetation) have been explained. In the second part, the most common platforms used on remote sensing (i.e. Landsat, Spot and IRS-1C), charecteristics of satellites and sensors have been given. In the following part, the digital image processing techniques covering rectification, enhancement (ratio imaging), classification (supervised and unsupervised) have been described. In the next part, Geographic Information System (GIS) has been explained. Its components, advantage, data types used in GIS have been given. In the application part of this study, the hazards in the forest area caused by the fyre burning were analised using the remotely sensed images. In order to determine the topographical structure of the area, standard topographic maps of 1/25000 scaled were used. The information about the structure of forest were derived by the forest map. With the digital image processing methods, such as spectral profile analysis, rationing, classification, the satellite images acquired after and before the fire were evaluated. From these remote sensing data spectral profile analysis, ratio images, classified images were obtained. Afterwards the images were rectified to the national projection system (UTM) using the ground control points to determine the type and the areal extent of the hazard. Finally all the data were added into the same database so as to construct a GIS. The aim of the establishing a GIS was to develop information and decision support systems to monitor and predict forest fire activity and to enhance fire management efficiency and effectiveness. In the result, differences occured between the ground data and the areak extent derived satellite image data were examined. In the last part the final evaluation of this application has been given. As shown in the final part the remote sensing data can be effectively used in analysis of forest fire damages. Especially GIS integration with remote sensing data is necessary and feasible to protect and manage forests before, during and after the fire burning.