Tuba Kurban

Tuba Kurban

Auto-generated profile, based on publicly available information

Tuba Kurban

Gazi Üniversitesi, Mühendislik Fakültesi

Research Focus & Metrics

Academic Career Length

20Years (2007-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
12 August 2009 in Sensors
Verified
Journal Article
Open access
PDF file
TKTuba Kurban
Tuba Kurban
EEErkan Esdok
Erkan Esdok
This paper introduces a comparison of training algorithms of radial basis function (RBF) neural networks for classification purposes. RBF networks provide effective solutions in many science and engineering fields. They are especially popular in the pattern classification and signal processing areas. Several algorithms have been proposed for training RBF networks. The Artificial Bee Colony (ABC) algorithm is a new, very simple and robust population based optimization algorithm that is inspired by the intelligent behavior of honey bee swarms. The training performance of the ABC algorithm is compared with the Genetic algorithm, Kalman filtering algorithm and gradient descent algorithm. In the experiments, not only well known classification problems from the UCI repository such as the Iris, Wine and Glass datasets have been used, but also an experimental setup is designed and inertial sensor based terrain classification for autonomous ground vehicles was also achieved. Experimental results show that the use of the ABC algorithm results in better learning than those of others.

1 October 2014 in Applied Soft Computing
Verified
Journal Article
Closed access
TKTuba Kurban
Tuba Kurban
PCPinar Civicioglu
Pinar Civicioglu
Rifat KurbanRifat Kurban
Rifat Kurban
EBErkan Besdok
Erkan Besdok

16 April 2021 in International Journal of Remote Sensing
Verified
Journal Article
Closed access
TKTuba Kurban
Tuba Kurban
Information provided from a single spectral band of a satellite image, may not be sufficient in most cases for classification, recognition and change detection applications. Therefore, bands with different spectral and spatial characteristics are combined to obtain a single fused image that contains complementary information. This study introduces a novel hybrid fusion method for remotely sensed infrared and visible images based on backtracking search algorithm (BSA) and Shearlet transform. Shearlet is an efficient processing method to transform spatial information and BSA is a powerful metaheuristic optimization method. Combining these techniques offers an efficient way to fuse low-resolution infrared and high-resolution panchromatic bands of satellite images. Extensive experiments proved that proposed method outperforms well-known multi-scale transforms based fusion methods in terms of both numerical and visual evaluations.

15 April 2022 in Geomatik
Verified
Journal Article
Closed access
AFAbdullah Fadhil
Abdullah Fadhil
TKTuba Kurban
Tuba Kurban
Kentsel yayılma küresel bir olgudur ancak Irak gibi gelişmekte olan ülkelerde çok daha hızlı bir şekilde gerçekleşmektedir. Kerkük, kentsel alanlarda hızla artan bir genişlemeye tanık olunan Irak kentlerinden biridir. Bu kentsel yayılma; hızlı nüfus artışı, plansız büyüme ve göç gibi çeşitli faktörlerden kaynaklanmaktadır. Bunun sonucunda kentsel ekosistem bu süreçten önemli ölçüde etkilenmektedir. Bu çalışmada, 2002-2018 yılları arasında Kerkük ilindeki arazi kullanımı/arazi örtüsü değişiklikleri CBS ve uzaktan algılama teknikleri kullanılarak incelenmiştir. Tespit edilen değişimler göstermektedir ki en büyük değişiklik %130 oranında bir artışla kentsel alanlarda gerçekleşmiştir. Bununla birlikte tarım arazilerinin %22 ve su alanlarının da %8 oranında genişlediği tespit edilmiştir. Buna karşılık, açık arazi alanında %3 oranında bir azalma olduğu da görülmüştür. Bu aşamadan sonra, gelecekte yaşanacak kentsel yayılma oranının öngörülmesi amacıyla karma bir Hücresel Otomata Markov Zincir modeli oluşturulmuştur. Model mevcut arazi kullanım haritası ile karşılaştırılmıştır ve Kappa istatistikleri kullanılarak doğrulanmıştır. Kappa standart, kappa konum ve kappa nicelik katsayıları sırasıyla 0.8799, 0.9143 ve 0.9154 olarak bulunmuştur. Bu sonuçlar, referans haritası ile karşılaştırma haritası arasında iyi bir örtüşme olduğunu göstermektedir. Ayrıca bu model, 2030 ve 2035 yıllarında gerçekleşmesi muhtemel kentsel yayılmanın tahmin edilmesi için kullanılmıştır.

1 April 2019
Verified
Conference Paper
Closed access
TKTuba Kurban
Tuba Kurban
EBErkan Besdok
Erkan Besdok
The ability to model an object or an environment using 3-dimensional point cloud data is very important for some areas such as photogrammetry, remote sensing, material processing and production, reverse engineering, construction industry, virtual reality and medicine. However, 3-dimensional point cloud data obtained with existing technologies contain some noise due to the nature of the measurement device and man-made errors. Therefore, it has great importance to filter raw point cloud data or surface elements derived from point cloud data to increase the quality of the 3D model. In this study, a filtering method was developed based on plane fitting by differential evolution algorithm to filter noisy point cloud data. The proposed heuristic algorithm-based filtering method is compared with the singular value decomposition method, which is frequently used in literature to obtain the plane parameters. Both visual and numerical results show that the plane fitting method based on the differential evolution algorithm is more successful to remove noise than the classical filtering method based on singular value decomposition.

1 May 2018
Verified
Conference Paper
Closed access
TKTuba Kurban
Tuba Kurban
EBErkan Besdok
Erkan Besdok
3-dimensional (3D) measurement and modeling, latterly has become an important issue to facilitate the analysis of a real object or an environment. However, obtained data at a time by existing methods such as laser scanning, structured light and stereo vision, cannot represent the entire object or the environment due to the reflections and limited field of view of the sensors. In order to generate a full 3D model, it is necessary to make measurements from different positions and locate all of them in a common coordinate system. This procedure called as point cloud registration. In this study, to registration of point clouds without any prior information, four basic affine transformation parameters are calculated by global optimization techniques such as Differential Evolution (DE) and Particle Swarm Optimization (PSO). Algorithm performances are compared with a fast, deterministic algorithm called Iterative Closest Point (ICP) that highly used in 3D registration literature. The studies show that, global optimization algorithms are quite successful in.