7 sonuçlar
Arama Sonuçları
Listeleniyor 1 - 7 / 7
Yayın An emprical point error model for TLS derived point clouds(International Society for Photogrammetry and Remote Sensing, 2016) Özendi, Mustafa; Akça, Mehmet Devrim; Topan, HüseyinThe random error pattern of point clouds has significant effect on the quality of final 3D model. The magnitude and distribution of random errors should be modelled numerically. This work aims at developing such an anisotropic point error model, specifically for the terrestrial laser scanner (TLS) acquired 3D point clouds. A priori precisions of basic TLS observations, which are the range, horizontal angle and vertical angle, are determined by predefined and practical measurement configurations, performed at real-world test environments. A priori precision of horizontal (??) and vertical (??) angles are constant for each point of a data set, and can directly be determined through the repetitive scanning of the same environment. In our practical tests, precisions of the horizontal and vertical angles were found as ??=±36.6 and ??=±17.8, respectively. On the other hand, a priori precision of the range observation (??) is assumed to be a function of range, incidence angle of the incoming laser ray, and reflectivity of object surface. Hence, it is a variable, and computed for each point individually by employing an empirically developed formula varying as ??=±2a'12 mm for a FARO Focus X330 laser scanner. This procedure was followed by the computation of error ellipsoids of each point using the law of variance-covariance propagation. The direction and size of the error ellipsoids were computed by the principal components transformation. The usability and feasibility of the model was investigated in real world scenarios. These investigations validated the suitability and practicality of the proposed method.Yayın Co-registration of 3d point clouds by using an errors-in-variables model(Copernicus Gesellschaft MBH, 2012-08-25) Aydar, Umut; Altan, Mehmet Orhan; Akyılmaz, Orhan; Akça, Mehmet DevrimCo-registration of point clouds of partially scanned objects is the first step of the 3D modeling workflow. The aim of co-registration is to merge the overlapping point clouds by estimating the spatial transformation parameters. In the literature, one of the most popular methods is the ICP (Iterative Closest Point) algorithm and its variants. There exist the 3D least squares (LS) matching methods as well. In most of the co-registration methods, the stochastic properties of the search surfaces are usually omitted. This omission is expected to be minor and does not disturb the solution vector significantly. However, the a posteriori covariance matrix will be affected by the neglected uncertainty of the function values. This causes deterioration in the realistic precision estimates. In order to overcome this limitation, we propose a new method where the stochastic properties of both (template and search) surfaces are considered under an errors-in-variables (EIV) model. The experiments have been carried out using a close range laser scanning data set and the results of the conventional and EIV types of the ICP matching methods have been compared.Yayın A theoretical comparison of ResNet and DenseNet architectures on the subject of shoreline extraction(Işık Üniversitesi, 2020-09-23) Ecevit, Mert İlhan; Çavdaroğlu, Gülsüm Çiğdem; Işık Üniversitesi, Fen Bilimleri Enstitüsü, Enformasyon Teknolojileri Yüksek Lisans ProgramıToday's Deep Learning technologies provides numerous approaches on the subject of convolutional networks. These approaches serve researchers to train datasets and generate wanted results from these datasets. Each CNN architecture has its own strong points and weak sides. Because of this situation a comparison between these architectures is a valuable asset. Image processing is a method that is frequently used to process remotely sensed data in remote sensing studies.. Between current architectures, RESNET and DENSENET architectures are chosen to be used by Dr. Çavdaroğlu for her project on TÜBİTAK. The result of this comparison will be used in that project in order to apply most ecient architecture. This thesis is written to draw outlines of RESNET and DENSENET and create a foresight for further projects which can be supported by this thesis. In order to achieve an accurate image recognition process in remote sensing domain, a preliminary research is requisite. As a research thesis this work serves the purpose of learning manner of works, performance indicators of RESNET and DENSENET convolutional networks. The result of this research will create a baseline for an academical project. At the other hand, comparison of these two convolutional network approaches provides information to decide which approach is more suitable for remote sensing projects depending upon the subject of the project. For future works on Remote Sensing this thesis work will serve a guideline and reason for preference. The presented thesis work has been developed as the technical feasibility of the 3501 TÜBITAK Project named "Uydu Görüntülerinden Kıyı Sınırlarının Derin Öğrenme Yöntemleri ile Otomatik Çıkarımı", applied by Dr. G. Çiğdem Çavdaroğlu, and the thesis results will be applied within the scope of the Project after the project acceptance.Yayın Spatial-Temporary analysis of Istanbul air pollution during the pandemic using Google Earth Engine and Google community mobility reports(Gök, Murat, 2023-06-30) Çavdaroğlu, Gülsüm Çiğdem; Arık, Ahmet OkanThe Covid-19 pandemic has brought drastic changes to people's daily life and environmental characteristics. To control the pandemic, all governments have implemented particular policies for their countries and imposed restrictions that affect people's daily life. The traffic index has decreased in many countries and cities depending on the restrictions. Therefore, restrictions in many countries and cities have positively impacted air quality. However, the opposite has also been observed in metropolitan cities. In this study, the change in the air quality of Istanbul, which is accepted as Turkey's largest metropolitan city, has been examined. First, the spatio-temporal distribution of air pollutants (NO2, CO, and SO2) has been analyzed using Sentinel-5P NRTI satellite images. Then six independent variable groups (traffic index of Istanbul, daily deaths in Istanbul, Google community mobility reports of Istanbul, fuel prices, stringency index of Turkey, two logical attributes regarding the Covid-19 restrictions and in-class education) were collected and combined to analyze the correlations between these variable groups and air pollutant concentrations. According to the spatial distribution graphs, there is a tendency to decrease NO2, CO, and SO2 pollutant concentrations in Istanbul when the restrictions are applied in Turkey. There was no significant relationship between the decrease in community mobility in Istanbul and pollutant concentrations, although an increase in air quality has been observed in many cities due to the restrictions of the Covid-19 pandemic.Yayın Correlation analysis between the community mobility and nighttime lights in the city of Istanbul, Turkey(2022) Çavdaroğlu, Gülsüm ÇiğdemThe COVID-19, which emerged in Wuhan, China, in 2019, has significantly affected people’s Daily lives, business environment, surrounding environment, and countries' economic and social conditions. This study aims to measure the correlation between community mobility changes in six different areas and nighttime lights in the city of Istanbul, Turkey. Nighttime light data used in the study was obtained from VIIRS Nighttime Day/Night Band Composites Version 1 using remote sensing methods via Google Earth Engine platform. Then the correlation between Nighttime light values and community mobility values was investigated. It has been observed that the correlation values have changed dramatically over the years. The most significant correlation values were observed for the year 2020. This is because 2020 is the year when the pandemic is most effective, and restrictions are at the highest level in Turkey. The increase in freedom in the following years caused a decrease in the correlation. When the correlation results covering the period of February 2020 - to January 2022 were examined, it was observed that there was no significant relationship between nighttime light values and Google Community Mobility Reports’ variables. Considering the correlation results for 2020, it was observed that there was a high negative correlation between nighttime light data and mobility trends for grocery and pharmacy, and mobility trends for places of work. In addition, there was a moderate negative correlation between nighttime light data and mobility trends for retail and recreation, and a moderate positive correlation between nighttime light and mobility trends for places of residence. When the correlation values of 2021 and the correlation values of the period 2021-2022 were examined, no significant relationship was observed.Yayın Google Earth engine based approach for finding fire locations and burned areas in Muğla, Turkey(Science Publishing Group, 2021-10-05) Çavdaroğlu, Gülsüm ÇiğdemForests are considered as one of the main sources of biodiversity. Forest fires caused by various reasons pose a high risk in terms of biodiversity. Therefore, mapping of fire zones is of great importance in determining the damage caused by the fire, managing the fire process, and planning the interventions in the fire zone. Although remote sensing is a fast and cost-effective methodology for mapping fire zones, the implementation of the remote sensing methodologies is problematic in some respects. The web-based Google Earth Engine makes possible to access the satellite imagery and process the imagery easily. The research area of this study is Muğla, Turkey in where many forest fires broke out in 2021 summer. This study provides an implementation of normalized burn ratio which is widely used to highlight burned areas on Google Earth Engine platform. Both vector data and satellite images were used in the study. The vector data is in the shape file format and was uploaded to the Google Earth Engine platform as a table. The Sentinel-2 imagery was used to calculate normalized burn ratio. The satellite imagery was clipped using the table data. The difference pre-fire and post-fire images was calculated, and the classes were assigned to the pixels according to the normalized burn ratio ranges. The study indicates that finding the burned areas and constructing the burn severity levels can be realized in 1.32 minutes on Google Earth Engine platform.Yayın İnsansız hava aracı ve Sentinel-2 görüntüleri kullanılarak ayçiçeği haritalamasına dayalı kovan yerleştirme karar destek sistemi(BZT Turan Publishing House, 2025-12-31) Yelmenoğlu, Elif Deniz; Aydın, Şahin; Çavdaroğlu, Gülsüm Çiğdem; Deniz, Hüseyin; Pajenado, Rex S.; Dilli, ŞirinAyçiçeği, yüksek nektar üretim kapasitesi nedeniyle gezici arıcılık faaliyetleri açısından stratejik öneme sahip tarımsal bitkilerden biridir. Ayçiçeği ekim alanlarının mekânsal ve zamansal dağılımı, arı kolonilerinin beslenme olanaklarını ve dolayısıyla bal verimini doğrudan etkilemektedir. Bu nedenle, arı kovanlarının uygun alanlara ve doğru zaman dilimlerinde yerleştirilmesi, gezici arıcılığın verimliliği açısından kritik bir karar sürecini oluşturmaktadır. Ancak mevcut uygulamalarda, kovan yer seçimi çoğunlukla arıcıların bireysel deneyimlerine ve sezgisel yaklaşımlarına dayalı olarak gerçekleştirilmekte; uzaktan algılama, görüntü işleme ve mekânsal analiz gibi veri temelli yöntemlerden yeterince yararlanılmamaktadır. Bu durum, potansiyel olarak verim kayıplarına ve kaynakların etkin kullanılmamasına yol açabilmektedir. Bu çalışmada, ayçiçeği yoğunluğunun doğru ve güvenilir biçimde belirlenmesi yoluyla kovan yerleştirme planlamasını desteklemeyi amaçlayan, çok ölçekli bir uzaktan algılama tabanlı karar destek çerçevesi önerilmektedir. Önerilen yaklaşım, saha ölçeğinde yüksek mekânsal çözünürlük sağlayan insansız hava aracı (İHA) görüntüleri ile bölgesel ölçekte geniş alan kapsama imkânı sunan Sentinel-2 uydu görüntülerinin entegrasyonuna dayanmaktadır. Çalışma alanı olarak, Türkiye’nin önemli ayçiçeği üretim merkezlerinden biri olan Kırklareli ili seçilmiş; veri seti, nektar üretiminin en yüksek olduğu ayçiçeği çiçeklenme dönemi dikkate alınarak oluşturulmuştur. Ayçiçeği tespiti, makine öğrenmesi tabanlı Random Forest sınıflandırma yöntemi kullanılarak gerçekleştirilmiş ve geliştirilen model %90,7 genel doğruluk değerine ulaşmıştır. Sınıf bazlı performans değerlendirmelerinde ise, ayçiçeği ekili alanlar ile ayçiçeği olmayan alanlar için F1-skoru her iki sınıf açısından da 0,91 olarak hesaplanmıştır. Bu sonuçlar, modelin hem nektar açısından zengin ayçiçeği alanlarını hem de ayçiçeği bulunmayan bölgeleri güçlü ve dengeli bir şekilde ayırt edebildiğini göstermektedir. Elde edilen ayçiçeği yoğunluk haritaları temel alınarak, ayçiçeği oranının yüksek olduğu alanlar arı kovanı yerleştirilmesi için uygun bölgeler olarak tanımlanmış; ayçiçeği yoğunluğunun düşük olduğu veya hiç bulunmadığı alanlar ise kovan yerleştirilmesine uygun olmayan bölgeler olarak değerlendirilmiştir. Çalışmadan elde edilen bulgular, çok ölçekli uzaktan algılama verilerinin makine öğrenmesi yöntemleriyle bütünleştirilmesinin, gezici arıcılık uygulamalarında veri temelli, güvenilir ve ölçeklenebilir karar destek sistemlerinin geliştirilmesine önemli katkılar sağlayabileceğini ortaya koymaktadır.
