6
Archived Theses
0
DOIs Assigned
0%
DOI Rate
Discipline
Kıyı metropolitan şehri Karachi'nin CBS tabanlıçok kriterli karar analizi ile hasar görebilirlik değerlendirmesi
The coastal city of Karachi is the financial capital of Pakistan while being the most populous in the country. It was originally inhabited as a fishing village, which later became an important port during British rule. Almost all the major infrastructure, key facilities, slums, and housing schemes are located on or near the coast. The city was not known to pose risk of any major disaster until recently. With about twenty million people living in the city, it makes the inhabitants exposed and vulnerable to hazards than ever before. Although the city lies close to some on-shore faults, but the main concern is about off-shore faults in Makran Subduction Zone off the Karachi coast. The city has been hit by a Tsunami during 1945 earthquake with wave height of 1.5 m. Furthermore, the city has chronic problem of urban flooding hazard, mostly due to monsoon rain, which often put city at standstill and cause causalities. Even though the mega city faces risk of several hazards, there has not been any significant research to understand the type or extent of hazards it is likely to face. In this study, vulnerability analysis of Karachi metropolitan is carried out by taking into account the disasters which affected the city in past. It includes both geophysical and climate-related hazards such as earthquake, tsunami, and flooding. Because of scarcity of data, methodology relies heavily on expert knowledge and judgement with selection and weight of hazard indicators are given using Delphi method. Hazard vulnerability assessments are realized by using GIS based multi criteria decision analysis (MCDA). Final vulnerability maps of Karachi obtained from Multi Criteria Decision Anlaysis reveal that the areas at South and West of Karachi near the shore with high population density are at grave danger against tsunami. Earthquake vulnerability map, however, shows that most of the areas, especially ones with high population density, households, and located on alluvial deposits, have either very high or high vulnerability to earthquake. That is because three parameters (geology, population density, and number of households) carry significant weights in earthquake vulnerability mapping. Lastly, flood hazard vulnerability map identifies district east, central, west, and parts of Korangi having high to very high vulnerability to flooding. These are generally the areas, which are often hit with floods by the rivers (Malir and Lyari), apart from being low altitude, and high population density. Keywords: Multi Hazards, Vulnerability Assessment, Multi Criteria Decision Analysis, Karachi, Analytical Hierarchy Process, Flood
Landsat 8 görüntü serisi kullanılarak fotosentetik pigment bollukları ile erken verim tahmini
Timely estimation of crop yields is critical for monitoring global food production by international organizations as well as governments, farmers and the private sector dealing with storage, import and export of crops and associated products. Satellite remote sensing has the capability to provide near real-time information on a global scale. Combining satellite data and soft computing techniques to predict crop yields is a very effective strategy for continually forecasting crop yields. This thesis presents a novel approach for accurate and sustainable estimation of crop yields based on estimated abundances of endmembers that may be attributed to photosynthetic pigments. Landsat 8 images acquired during the time of the phenological cycle when plants have maximum greenness are the inputs to find endmembers and abundances within the pure wheat crop pixels using Robust Collaborative Nonnegative Matrix Factorization (R-CoNMF) unmixing algorithm. The endmembers are optimized to maximize the predictive power of the abundances for the yields. Wheat yields were then estimated with the four abundances, their relevant interactions, ten important agrometeorological parameters, including parameters proposed in this thesis for the first time, and four different vegetation indices using three different machine learning algorithms (Generalized Linear Model (GLM), Artificial Neural Network (ANN) and Random Forest (RF)). Harvester records from 142 wheat fields distributed in 31 provinces of Turkey were used as the ground truth for testing the algorithm. In the literature, the coefficient of determination (R2) is used as a proxy to show how good the relationship is between the estimated and real figures. According to these calculations, the yields were estimated with 64% accuracy when only the abundances were used in the GLM algorithm, 78% accuracy when ANN was used for yield estimation and 82% accuracy was reached when applying RF to all of the parameters. The similarity of the endmembers to photosynthetic pigment spectral signatures along with their predictive power suggested their relevance to the pigments. Although the R-CoNMF algorithm performs a linear unmixing of the intimate mixture of the photosynthetic pigments, the interactions of the abundances used in the endmember optimization and in classifications partially handle the non-linearity using the bilinear model. These results can be considered as a great success when using multispectral satellite data only and are recognized as a clear indication that much better results would be achieved while using images from future hyperspectral space missions like HyspIRI.
Akıllı kart verilerinin kullanılarak toplu ulaşım güvenilirlik ölçütlerinin elde edilmesi: Konya şehrinden 2 örnek hat çalışması
In Intelligent Transportation Systems (ITS) applications for Public Transit (PT), it is almost customary to monitor and collect bus trajectory data in addition to smart card systems, which includes geocoded information on ticketing. These datasets enable system managers to derive PT reliability and Level-of-Service (LOS) measures, such as average travel time, dwell time, etc, at a PT line level. However, it is also necessary to derive such measures spatio-temporally to get PT characteristics in metropolitan regions with heterogeneous land use and congestion patterns. This study focuses on processing and evaluation of smart card data (SCD) to determine PT reliability, using the case study results from Konya PT smart card system. A major goal of the study was to convert geocoded SCD to create time-space diagrams using linear referencing to a bus route. A set of data preprocessing studies have been conducted by using a database management software and a GIS software (PostgreSQL and ArcGIS) for converting data. Afterward, GIS-based technics known as Linear Referencing (LR) and Dynamic Segmentation (DynSeg) have been implemented to BTD coordinates gathered from public BTD in ArcGIS software to obtain the kilometers (aka Station Km) of the smart card records on the relevant route segments. Arrival times to the bus stops have been estimated, which also allowed estimation of inter-stop travel times (ISTT) along a route. Arrival time estimations (ATEs) have been used to visualize bus trajectory with time-space diagrams and to calculate total travel times and Travel Time (TT) reliability indicators. As a result, a route based evaluation has been made over two study PT lines in the Konya PT network by using these calculated TT reliability indicators and time-space diagrams.
Çok-zamanlı çok-bantlı uydu görüntüleri kullanarak Akdeniz çalı sınıflandırması
Shrublands, which have a crucial role in retaining the ecological balance, constitute an important part of the Mediterranean ecosystems. However, their composition, distribution and dynamics are not well understood. It is necessary to know the distribution of the alliances at regional scale in order to construct models that explain their dynamics. Such models will help researchers to evaluate their role in ecosystems and predict their responses to climate change. Necessary alliance distribution maps can only be produced by employing remote sensing techniques. This study presents a methodology that generates alliance-level woodland/shrubland maps of the Mediterranean region in southern Turkey from satellite images using various machine learning techniques with different parameter combinations. Multi-temporal images are used to extract information from vegetation phenology. Topographic and meteorological data are also used for improving classification. Cross-validation is performed using a ground-truth data set of 7452 polygons. Results show that detailed and accurate maquis shrubland classification is possible using a combination of environmental features and multi-spectral and multi-temporal satellite images. Addition of the environmental features to remotely sensed ones improved classification accuracy by 16%. The Random Forest (RF) algorithm is found to improve classification accuracy by 35.9% and 13.9% relative to Support Vector Machine and Quadratic Discriminant Analysis algorithms, respectively. Alliance-level classification maps of maquis acquired from RF classification are produced with 64.0-82.1% overall accuracy. Large-scale shrub classification method will have important implications on natural resource management and other ecological applications.
Alpin ağaç sınırı ekotonunun landsat tm görüntüleri kullanılarak belirlenmesi ve mekansal-zamansal analizi
Alpine treeline ecotone (ATE) is the transition zones between forests and alpine grasslands. Because of its ecological importance due to its unique biodiversity, understanding the characteristics of the transition zone is essential. Recent climate change research has shown that the ATE tends to shift upwards. Understanding this upward shift enables the development of climate change indicators for mountain ecosystems and provides insight into efforts towards improving adaptation and mitigation measures. This thesis aims to develop and apply a methodology for objectively defining and delineating a tree line that is also ecologically meaningful. Another aim is to reveal the shift in ATE for a study area in which this altitudinal shift is expected to have been substantial in the last decades. Within this context, the factors that determine the spatial configuration of the ATE have been investigated through the use of remotely sensed resources. The study area is in the Western Taurus Mountains, located in the Mediterranean region of Turkey. An algorithm with four steps has been developed to delineate treeline for any given time during the study period, using Landsat images of the relevant years. For each step, the methodology was chosen with the aim of minimising the need for human interpretation as much as possible and for developing a reproducible method. These steps include obtaining cloud-free seasonal composites from Landsat images, determining tree percentages over the area through spectro-temporal unmixing, and characterising the transition of ATE through fitting a sigmoid response to tree percentages along uphill transects, and modelling of the ATE transition using Random Forest Regression. Applying this algorithm has revealed that topographical variables combined with information on the percentage of canopy cover can be used effectively while modelling treeline ecotones. Outcomes of the model indicate a downward shift of the treeline on west face of the Dedegöl Mountain for some slopes since 1984, against theoretical expectations or contrary to observations of increases elsewhere in the world. However, on eastern slopes, the shift is indicated to be upwards. This study can provide input for estimating future shifts in ATE and for further development of models for various climates and latitudes. Also, the algorithm can easily be adapted to other satellite data, thus enabling higher resolution outcomes.
Yüksek çözünürlüklü nokta bulutlarından çoklu çözünürlüklü düzlemsellik tabanlı yaklaşım kullanarak dijital arazi modeli çıkarımı
Digital Elevation Model (DEM) is a mathematical representation of the elevation of the Earth's surface. There are two types of DEM, namely Digital Surface Model (DSM) and Digital Terrain Model (DTM). DSM contains natural (bare-ground, trees, bushes, etc.) and artificial above-ground objects (buildings, vehicles, powerlines, etc.), while DTM covers only the bare earth without anything on it. Above-ground objects need to be removed to extract the DTM, which is a tedious task. This thesis proposes an algorithm that extracts DTM from aerial point clouds using a robust multi-resolution planarity-based divide-and-conquer algorithm. In this approach, the problem is handled in few simple steps rather than trying to solve the problem at once. The approach contains different planarity checks to get rid of nonplanar above-ground objects, segmentation step to find rough ground points, and an interpolation step to obtain the final DTM. In this thesis, ground points are assumed planar, and planar patches are detected as ground candidates. First, approximate planarity values are calculated by using neighboring points. This helps to eliminate most of the above-ground objects such as vehicles, trees, posts, etc. Nevertheless, since the building facades and roofs are also planar, a second planarity check is needed in different resolutions. For this purpose, the grid planarity values are checked. The grids that do not fit a plane within the given threshold are marked as nonplanar. The second planarity check helps to get rid of the building facades and the vertical planes. After removing building facades, getting benefit from the sparsity between ground candidates and the roof points, a region growing segmentation is utilized to segment the remaining ground candidates for rough ground surface calculation. The segments far from the rough ground surface are omitted. By doing so, the roof points can be eliminated. Lastly, the ground points are interpolated to obtain the resulting DTM raster. Although the input point cloud is already classified as ground and non-ground, it has some errors. The input point cloud is used to create a DTM; then, the resulting DTM is manually edited to use it as a ground truth. The accuracy assessment is done on interpolated DTM rasters. Using a manually corrected ground truth, Root Mean Square Error (RMSE) is calculated for two datasets with different characteristics having 1.00 m and 2.20 m spatial resolutions. The results are compared with two existing DTM extraction algorithms and RMSE values are close to these solutions. The RMSE values are 0.25 m and 0.70 m, respectively. Results indicate that an accurate DTM extraction is possible using a combination of only planarity values. Keywords: Digital terrain model, digital elevation model, planarity, covariance features, principal component analysis, point cloud