Tree detection and counting have been performed using conventional methods or high costly remote ... more Tree detection and counting have been performed using conventional methods or high costly remote sensing data. In the past few years, deep learning techniques have gained significant progress in the remote sensing area. Namely, convolutional neural networks (CNNs) have been recognized as one of the most successful and widely used deep learning approaches and they have been used for object detection. In this paper, we employed a Mask R-CNN model and feature pyramid network (FPN) for tree extraction from high-resolution RGB unmanned aerial vehicle (UAV) data. The main aim of this paper is to explore the employed method in images with different scales and tree contents. For this purpose, UAV images from two different areas were acquired and three big-scale test images were created for experimental analysis and accuracy assessment. According to the accuracy analyses, despite the scale and the content changes, the proposed model maintains its detection accuracy to a large extent. To our knowledge, this is the first time a Mask R-CNN model with FPN has been used with UAV data for tree extraction.
International journal of environment and geoinformatics, Mar 19, 2023
Rapid population growth, natural events, and increasing industrialization are among the factors a... more Rapid population growth, natural events, and increasing industrialization are among the factors affecting land use. To keep this change under control and to make sound plans, it is necessary to control the changes. In this study, the spatial use change in the Eskişehir region between the years 1990-2018 was examined with CORINE data. Based on this determined change, an urban change model was created with the multivariate regression method. As a result of the evaluations, while an increase was observed in urban areas and pastures between 1990-2018, a decrease was determined in agricultural and forest areas. This change is defined as 43.74% in urban areas, 3.28% in agricultural areas, 7.78% in forest areas, and 60.10% in pasture areas. SMOReg, MLP Regressor, and M5P Model Tree methods were used for the estimation study to be carried out with the obtained spatial change data. Urban values for 2018 were estimated to find the best method. Finally, the areas of 2030 were estimated with the method that gave the best results. The results demonstrated the usability of modeling using CORINE data.
Photogrammetric Engineering and Remote Sensing, Feb 1, 2023
Monitoring trees is necessary to manage and take inventory of forests, monitor plants in urban ar... more Monitoring trees is necessary to manage and take inventory of forests, monitor plants in urban areas, distribute vegetation, monitor change, and establish sensitive and renewable agricultural systems. This study aims to automatically detect, count, and map apricot trees in an orthophoto, covering an area of approximately 48 ha on the ground surface using two different algorithms based on deep learning. Here, Mask region-based convolutional neural network (Mask R-CNN) and U-Net models were run together with a dilation operator to detect apricot trees in UAV images, and the performances of the models were compared. Results show that Mask R-CNN operated in this way performs better in tree detection, counting, and mapping tasks compared to U-Net. Mask R-CNN with the dilation operator achieved a precision of 98.7%, recall of 99.7%, F1 score of 99.1%, and intersection over union (IoU) of 74.8% for the test orthophoto. U-Net, on the other hand, has achieved a recall of 93.3%, precision of 97.2%, F1 score of 95.2%, and IoU of 58.3% when run with the dilation operator. Mask R-CNN was able to produce successful results in challenging areas. U-Net, on the other hand, showed a tendency to overlook existing trees rather than generate false alarms.
Environmental Monitoring and Assessment, Sep 10, 2022
the manufacturing industry, and the use of fossil fuels in industrial and residential activities ... more the manufacturing industry, and the use of fossil fuels in industrial and residential activities (Angelevska et al., 2021; Ghasempour et al., 2021). Carbon monoxide (CO), nitrogen dioxide (NO 2), ozone (O 3), sulfur dioxide (SO 2), and particulate matter (including PM 10 and PM 2.5) emissions from human and natural sources have a substantial influence on individual health and well-being (Gopalakrishnan et al., 2018). With more than half of the world's population living in cities, the impact of air pollution on public health must be addressed. Energy consumption, industrial emissions, and automobile traffic all rise when cities develop in population and size, all of which can have a negative impact on air quality (Kahyaoğlu-Koračin et al., 2009). Conversion of forests, grasslands, and cropland to urban development, industrial complexes, and big commercial areas frequently results in increased emissions. Urban sprawl is the most severe example of this sort of growth, characterized by dispersed patterns of low-density development, which is frequently automobile-oriented (Superczynski & Christopher, 2011). Inevitably, air quality varies based on land cover and as the environment changes. The air quality impacts of different land covers have been researched in a number of studies. Thus, the air quality impacts of grasslands and shrublands have been investigated in the USA (Gopalakrishnan et al., 2018) for the primary purpose of estimating the pollution removal capacity of canopy cover on a national level. Spatial interpolation of air pollution measurements was done, and the land cover was considered (Janssen et al., 2008). As a land use Abstract With the increased urbanization, the rise of the manufacturing industry, and the use of fossil fuels, poor air quality is one of the most serious and pressing problems worldwide. The COVID-19 outbreak prompted absolute lockdowns in the majority of countries throughout the world, posing new research questions. The study's goals were to analyze air and temperature parameters in Turkey across various land cover classes and to investigate the correlation between air and temperature. For that purpose, remote sensing data from MODIS and Sentinel-5P TROPOMI were used from 2019 to 2021 over Turkey. A large amount of data was processed and analyzed in Google Earth Engine (GEE). Results showed a significant decrease in NO 2 in urban areas. The findings can be used in long-term strategies for lowering global air pollution. Future research should look at similar investigations in various study sites and evaluate changes in air metrics over additional classes.
International journal of environment and geoinformatics, Apr 12, 2019
Remote sensing technologies provide very important big data to various science areas such as risk... more Remote sensing technologies provide very important big data to various science areas such as risk identification, damage detection and prevention studies. However, the classification processes used to create thematic maps to interpret this data can be ineffective due to the wide range of properties that these images provide. At this point, there arises a requirement to optimize the data. The first objective of this study is to evaluate the performance of the Bat Search Algorithm which has not previously been used for improving the classification accuracy of remotely sensed images by optimizing attributes. The second objective is to compare the performance of the Genetic Algorithm, Bat Search Algorithm, Cuckoo Search Algorithm and Particle Swarm Optimization Algorithm, which are used in many areas of the literature for the optimization of the attributes of remotely sensed images. For these purposes, an image from the Landsat 8 satellite is used. The performance of the algorithms is compared by classifying the image using the K-Means method. The analysis shows a 10-22% increase in overall accuracy with the addition of attribute optimization.
Anadolu Üniversitesi bilim ve teknoloji dergisi -b- teorik bilimler, Dec 31, 2018
Binalar deprem afetinden en fazla etkilenen nesnelerdir. Deprem sonrası yıkılan binaların tespit ... more Binalar deprem afetinden en fazla etkilenen nesnelerdir. Deprem sonrası yıkılan binaların tespit edilmesi, hem mevcut durumunun belirlenmesi hem de hızlı müdahale açısından önemlidir. Son yıllarda gelişen insansız hava araçları, üzerlerine takılan kamera sistemleri sayesinde yeryüzüne ait çok yüksek çözünürlüklü görüntüler elde edilebilmektedir. Bu görüntülerden üretilen ürünler aracılığı ile istenilen amaca yönelik bilgiler çıkarılabilmektedir. Bu çalışmada, 2015 ve 2014 yıllarında insansız hava aracı ile yüksek çözünürlüklü görüntüleri elde edilen bir alanda, yıkılan binaların tespiti gerçekleştirilmiştir. Bina tespiti işlemi senaryo bir olay üzerinden yapılmıştır. Bu kapsamda, 2015 yılı görüntüleri deprem öncesi, 2014 yılı görüntüleri deprem sonrası olarak ele alınmıştır. Her iki yıla ait görüntüler işlenerek alana ait sayısal yükseklik modeli ve ortofoto görüntü üretilmiştir. Üretilen bu verilere nesne tabanlı sınıflandırma işlemi uygulanarak, çalışma alanında yer alan binalar çıkarılmıştır. Her iki yıla ait bina sınıflarının karşılaştırılması ile 2015 yılında alanda mevcut olup, 2014 yılında alanda olmayan 11 bina başarı ile tespit edilmiştir.
Forest fires cause a lot of damage in Turkey, due to the geographical location, high temperatures... more Forest fires cause a lot of damage in Turkey, due to the geographical location, high temperatures, and low humidity levels. Accurate determination of burned forest areas is crucial for correct damage assessment studies, fire risk calculations, and review of the forest regeneration processes. In this study, we compare the performances of unsupervised classification methods (which have not been used to map burned areas before) of burned area extraction from medium resolution satellite images with K-means. In this regard, the areas affected by fire in the Kumluca and Adrasan regions in 2016, Alanya and Gümüldür regions in 2017 and Athens region in 2018 are determined using Landsat 8 images. For this purpose, Canopy, M-tree, a hierarchical clustering algorithm, and a learning vector quantization which are frequently used in the literature are applied to determine the burned area, and the results obtained are compared with the results obtained from K-means. The results show that unsupervised classification methods can be used to map burned areas. The hierarchical clustering and K-means algorithms provide the most accurate results in mapping burned areas in most of the regions used in the study.
Değişim tespitinin temel amacı, aynı bölgenin farklı zamanlarda çekilmiş görüntülerini karşılaştı... more Değişim tespitinin temel amacı, aynı bölgenin farklı zamanlarda çekilmiş görüntülerini karşılaştırarak her piksele değişim olan ve olmayan olmak üzere ikili kodlanmış etiketler atamaktır. Yüksek çözünürlüklü optik uzaktan algılama görüntülerine dayalı değişim tespiti çalışma alanında bulunan nesnelerin karmaşıklığı ve iki tarih arasındaki farklı görüntüleme koşulları nedeniyle zorlu bir görevdir. Bu çalışmada LEarning, VIsion and Remote sensing (LEVIR)-CD veri seti ile eğitilmiş, derin sinir ağı temelli Bitemporal Görüntü Dönüştürücü (Bitemporal Image Transformer-BIT) ve STANet modellerinin farklı çalışma alanlarındaki performansının araştırılması amaçlanmıştır. Elde edilen sonuçlar LEVIR-CD veri seti ile eğitilmiş olan BIT ve STANet modellerinin yüksek doğruluk ile değişim tespiti gerçekleştirmesi için ek eğitim veri setine ihtiyaç duyduğunu göstermektedir.
Süper çözünürlük, çeşitli yollarla görüntü çözünürlüğünü artırmayı amaçlayan ve son yıllarda deri... more Süper çözünürlük, çeşitli yollarla görüntü çözünürlüğünü artırmayı amaçlayan ve son yıllarda derin öğrenme alanındaki gelişmelerle beraber daha iyi sonuçların elde edildiği bir görüntü iyileştirme yöntemidir. Keşif-gözetleme, nesne tespiti, çeşitli tıp uygulamaları gibi birçok alanda kendine uygulama sahası bulan bu yöntemle düşük çözünürlüklü görüntülerden daha fazla bilgi çıkarımı yapmak mümkün hale gelmektedir. Görüntü iyileştirme işlemi geleneksel yollarda interpolasyon gibi matematiksel tahminleme yöntemleriyle başarılmaya çalışılırken derin öğrenmeye dayalı modeller bunu çeşitli Konvolüsyonel Sinir Ağı mimarilerini etiketlenmiş verilerle beraber kullanarak gerçekleştirmektedir. Bu çalışmada yaklaşık 3 metre yersel çözünürlüğe sahip 8-bitlik PlanetScope uydu görüntüleri biri geleneksel bikübik interpolasyon, diğeri derin öğrenmeye dayalı ESPCN (Efficient Sub-Pixel Convolutional Neural Network) modelleri kullanılarak iyileştirilmiş ve sonuçlar PSNR (Peak Signal to Noise Ratio) değerleri cinsinden karşılaştırılmıştır. ESPCN'nin eğitiminde farklı veri setleri kullanılmış ve neticeye etkisi gözlenmiştir. Elde edilen sonuçlara göre; eğitimde veri sayısını artırmanın ve eğitim verisi türünün test verisiyle benzer olmasının sonuçları olumlu etkilediği ve ayrıca her ne kadar farklı türde ve az sayıda veriyle eğitilmiş olsa bile ESPCN yönteminin geleneksel bikübik interpolasyona göre daha iyi sonuçlar ürettiği ortaya konulmuştur.
Su kaynakları yaşamın devamlılığında önemli bir rol oynamaktadır. Bu nedenle su kaynaklarının har... more Su kaynakları yaşamın devamlılığında önemli bir rol oynamaktadır. Bu nedenle su kaynaklarının haritalanması ve değişikliklerin izlenmesi gerekmektedir. Su kaynaklarının izlenmesi, kontrolü ve koruma çalışmalarında uzaktan algılama teknolojileri önemli veriler sağlamaktadır. Bu veriler, su kütleleri ile ilgili çalışmalarda planlayıcılar için önemlidir. Bu çalışmada Manisa'ya 70 km uzaklıkta bulunan Gölmarmara ilçesinde yer alan Marmara Gölü su yüzeyinin değişim analizi gerçekleştirilmiştir. Ek olarak Marmara Gölünün gelecekteki alansal değişimine ait tahminleme çalışması gerçekleştirilmiştir. Bu doğrultuda yüzey alanları, çalışma alanına ait 2002-2021 yıllarına ait Landsat 7 görüntülerinin kontrolsüz sınıflandırma yöntemi ile analizi sonucunda elde edilmiştir. Bunun yanında alana ait yağış, sıcaklık ve arazi yüzey sıcaklığı (LST) verileri Google Earth Engine yardımıyla elde edilmiştir. Elde edilen veriler kullanılarak en doğru tahminlemeyi yapabilmek amacıyla Radyal Tabanlı Fonksiyon (RBF Regressor), Doğrusal Regresyon (Lineer Regression), Toplamsal Regresyon (Additive Regression) ve Çok Katmanlı Perceptron Sınıflandırıcı (MultiLayer Perceptron Classifier) yöntemleri kullanılmıştır. 2002-2012 yılları arasındaki veriler kullanılarak 2013 ve 2021 yılları arasındaki değişim belirlenmiştir. Sonuçlar incelendiğinde en iyi tahminin R2= 0.91 ile Çok Katmanlı Perceptron CS ile elde edildiği gözlemlenmiştir. Bu yöntem ile 2022 ve 2026 yılları için gerçekleştirilen tahmin çalışması sonucunda gölün çok daha fazla küçüleceği ve 1.56 km2' ye ulaşacağı öngörülmüştür.
This article is an open access article distributed under the terms and conditions of the Creative... more This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY
Computers, Environment and Urban Systems, Jul 1, 2018
Geocoding is a tool that can be used in many areas such as the development of disaster prevention... more Geocoding is a tool that can be used in many areas such as the development of disaster prevention systems, crime mapping and the monitoring of communicable diseases, and which has gradually gained importance. However, the use of geocoding is not yet possible in some areas where it could serve as an effective tool, for various reasons such as inconsistencies in address formats, including inaccurate numbering systems, misspellings, the use of abbreviations and a lack of data that refers to the geocoding process. This study seeks to address these problems by way of a standardization process. To that end, it employs a method that decomposes addresses used as input data in geocoding, identifies spelling mistakes and abbreviations, and reorganizes the addresses through the Natural Language Process (NLP). As test data, the addresses of primary schools in the district of Eskisehir are taken. First the geocoding process is performed on the data set, using both Google geocoding API and ArcGIS geocoding API. Then, the addresses are reformatted into three address formats by applying standardization processes. Geocoding is performed on the re-formatted addresses and the results compared to the non-standardized results. The standardization used is shown to make a significant improvement in the accuracy of the geocoding results. The method used in this study is significant not only in increasing the accuracy of the geocoding process, but also in sustaining its wider use.
Su kaynakları yaşamın devamlılığında önemli bir rol oynamaktadır. Bu nedenle su kaynaklarının har... more Su kaynakları yaşamın devamlılığında önemli bir rol oynamaktadır. Bu nedenle su kaynaklarının haritalanması ve değişikliklerin izlenmesi gerekmektedir. Su kaynaklarının izlenmesi, kontrolü ve koruma çalışmalarında uzaktan algılama teknolojileri önemli veriler sağlamaktadır. Bu veriler, su kütleleri ile ilgili çalışmalarda planlayıcılar için önemlidir. Bu çalışmada Manisa'ya 70 km uzaklıkta bulunan Gölmarmara ilçesinde yer alan Marmara Gölü su yüzeyinin değişim analizi gerçekleştirilmiştir. Ek olarak Marmara Gölünün gelecekteki alansal değişimine ait tahminleme çalışması gerçekleştirilmiştir. Bu doğrultuda yüzey alanları, çalışma alanına ait 2002-2021 yıllarına ait Landsat 7 görüntülerinin kontrolsüz sınıflandırma yöntemi ile analizi sonucunda elde edilmiştir. Bunun yanında alana ait yağış, sıcaklık ve arazi yüzey sıcaklığı (LST) verileri Google Earth Engine yardımıyla elde edilmiştir. Elde edilen veriler kullanılarak en doğru tahminlemeyi yapabilmek amacıyla Radyal Tabanlı Fonksiyon (RBF Regressor), Doğrusal Regresyon (Lineer Regression), Toplamsal Regresyon (Additive Regression) ve Çok Katmanlı Perceptron Sınıflandırıcı (MultiLayer Perceptron Classifier) yöntemleri kullanılmıştır. 2002-2012 yılları arasındaki veriler kullanılarak 2013 ve 2021 yılları arasındaki değişim belirlenmiştir. Sonuçlar incelendiğinde en iyi tahminin R2= 0.91 ile Çok Katmanlı Perceptron CS ile elde edildiği gözlemlenmiştir. Bu yöntem ile 2022 ve 2026 yılları için gerçekleştirilen tahmin çalışması sonucunda gölün çok daha fazla küçüleceği ve 1.56 km2' ye ulaşacağı öngörülmüştür.
International Journal of Engineering and Geosciences
Remote sensing Google Earth Engine Forest areas changes Statistical Analyze Forest area losses ar... more Remote sensing Google Earth Engine Forest areas changes Statistical Analyze Forest area losses are one of the most significant changes in land cover. These losses negatively affect ecosystems and cause severe economic and social life problems. It is necessary to monitor the process carefully and analyze the effects well to minimize all these negative effects in forest land losses and improve the development in urban areas positively. It is of great importance that these analyses are carried out quickly and accurately in terms of developing the natural environment. In this study, the effects that cause forest losses in the Mediterranean Region over the years are examined with the data obtained with the Google Earth Engine (GEE). Within the scope of the study, the changes in forest areas in the Mediterranean Region between 2004 and 2019 have been examined by considering many factors. In the study, Normalized Difference Vegetation Index (NDVI), precipitation, temperature, land surface temperature, aerosol optical depth, ozone, fire, urban areas, and population data were obtained with GEE. The data obtained were analyzed statistically, and the factors affecting the losses in forest areas the most were determined.
With the developing technology and automation, automatic labelling of images is of great importan... more With the developing technology and automation, automatic labelling of images is of great importance for automatic mapping. However, the most significant disadvantage of this method is that the classes’ labels cannot be generated automatically. In the current remote sensing literature, understanding and automatically labelling clusters obtained from the clustering process without a training phase is a problem that requires effective solutions. In this study, in order to solve this problem, we present a methodology that creates labels without any training phase. We use the bands in the image and Corine data in this process. The methodology uses a database created by examining the spectral characteristics of land classes from sample images collected from various geographies and time periods. The spectral index values of the unlabelled classes obtained are evaluated using this database, and the relevant label is assigned to each class. This database was created by analyzing Sentinel-2 Level-1 images of the Mediterranean and the Black Sea regions in Turkey. Then, these labels compare with the Corine classes corresponds to each pixel according to the ruleset. This developed approach aims to automatically label land, a green agricultural area, forest, urban area, and uncultivated agricultural area. The reason for choosing these areas is that they are the areas that generally make up the environment and a large part of the ecosystem, which are important areas that many researchers frequently use in their studies. The methodology developed was tested with Sentinel 2 images of Gemlik, Hatay regions from Turkey, and Agioi Apostoli region from Greece. The results of the accuracy analysis are 80%, 83%, and %82 for Gemlik, Hatay, and Agioi Apostoli areas.
Journal of the Indian Society of Remote Sensing, 2022
Urban areas of major cities in developing countries are expanding rapidly due to rapid population... more Urban areas of major cities in developing countries are expanding rapidly due to rapid population growth and industrialization. Determining the reasons for rapid urbanization and the amount of urban area that will be needed in the future is important for the planned growth of these cities. In this study, the LU/LC change of Eskisehir city center that has a rapid rate of urbanization and industrialization in Turkey between 1984 and 2020 was determined. Then, the area needed for urban and industrial areas in 2030 was investigated using the Cullullar Automata–Markov Chain (CA–MC) hybrid model simulation. Each year's LU/LC map was produced with over 80% kappa accuracies performing the pixel-based Random Forest (RF) algorithm for change analysis. In the change analysis made between 1984 and 2020, it was determined that there was a change of 117% in the urban area and 977% in the industrial area. The prediction was made for the validation of the CA–MC model from the period 2000–2010 to 2020. When compared with the 2020 LU/LC map, the model success was obtained as 0.84, 0.87, and 0.87 in the Kstandard, Klocation, and Kno Kappa metrics, respectively. By using the 2010–2020 periods in the estimation of 2030, it has been observed that the urban and industrial area will increase by 22.95%; therefore, there will be decreases in agriculture and other natural areas as in previous years.
Monitoring trees is necessary to manage and take inventory of forests, monitor plants in urban ar... more Monitoring trees is necessary to manage and take inventory of forests, monitor plants in urban areas, distribute vegetation, monitor change, and establish sensitive and renewable agricultural systems. This study aims to automatically detect, count, and map apricot trees in an orthophoto, covering an area of approximately 48 ha on the ground surface using two different algorithms based on deep learning. Here, Mask region-based convolutional neural network (Mask R-CNN) and U-Net models were run together with a dilation operator to detect apricot trees in UAV images, and the performances of the models were compared. Results show that Mask R-CNN operated in this way performs better in tree detection, counting, and mapping tasks compared to U-Net. Mask R-CNN with the dilation operator achieved a precision of 98.7%, recall of 99.7%, F1 score of 99.1%, and intersection over union (IoU) of 74.8% for the test orthophoto. U-Net, on the other hand, has achieved a recall of 93.3%, precision of ...
Tree detection and counting have been performed using conventional methods or high costly remote ... more Tree detection and counting have been performed using conventional methods or high costly remote sensing data. In the past few years, deep learning techniques have gained significant progress in the remote sensing area. Namely, convolutional neural networks (CNNs) have been recognized as one of the most successful and widely used deep learning approaches and they have been used for object detection. In this paper, we employed a Mask R-CNN model and feature pyramid network (FPN) for tree extraction from high-resolution RGB unmanned aerial vehicle (UAV) data. The main aim of this paper is to explore the employed method in images with different scales and tree contents. For this purpose, UAV images from two different areas were acquired and three big-scale test images were created for experimental analysis and accuracy assessment. According to the accuracy analyses, despite the scale and the content changes, the proposed model maintains its detection accuracy to a large extent. To our knowledge, this is the first time a Mask R-CNN model with FPN has been used with UAV data for tree extraction.
International journal of environment and geoinformatics, Mar 19, 2023
Rapid population growth, natural events, and increasing industrialization are among the factors a... more Rapid population growth, natural events, and increasing industrialization are among the factors affecting land use. To keep this change under control and to make sound plans, it is necessary to control the changes. In this study, the spatial use change in the Eskişehir region between the years 1990-2018 was examined with CORINE data. Based on this determined change, an urban change model was created with the multivariate regression method. As a result of the evaluations, while an increase was observed in urban areas and pastures between 1990-2018, a decrease was determined in agricultural and forest areas. This change is defined as 43.74% in urban areas, 3.28% in agricultural areas, 7.78% in forest areas, and 60.10% in pasture areas. SMOReg, MLP Regressor, and M5P Model Tree methods were used for the estimation study to be carried out with the obtained spatial change data. Urban values for 2018 were estimated to find the best method. Finally, the areas of 2030 were estimated with the method that gave the best results. The results demonstrated the usability of modeling using CORINE data.
Photogrammetric Engineering and Remote Sensing, Feb 1, 2023
Monitoring trees is necessary to manage and take inventory of forests, monitor plants in urban ar... more Monitoring trees is necessary to manage and take inventory of forests, monitor plants in urban areas, distribute vegetation, monitor change, and establish sensitive and renewable agricultural systems. This study aims to automatically detect, count, and map apricot trees in an orthophoto, covering an area of approximately 48 ha on the ground surface using two different algorithms based on deep learning. Here, Mask region-based convolutional neural network (Mask R-CNN) and U-Net models were run together with a dilation operator to detect apricot trees in UAV images, and the performances of the models were compared. Results show that Mask R-CNN operated in this way performs better in tree detection, counting, and mapping tasks compared to U-Net. Mask R-CNN with the dilation operator achieved a precision of 98.7%, recall of 99.7%, F1 score of 99.1%, and intersection over union (IoU) of 74.8% for the test orthophoto. U-Net, on the other hand, has achieved a recall of 93.3%, precision of 97.2%, F1 score of 95.2%, and IoU of 58.3% when run with the dilation operator. Mask R-CNN was able to produce successful results in challenging areas. U-Net, on the other hand, showed a tendency to overlook existing trees rather than generate false alarms.
Environmental Monitoring and Assessment, Sep 10, 2022
the manufacturing industry, and the use of fossil fuels in industrial and residential activities ... more the manufacturing industry, and the use of fossil fuels in industrial and residential activities (Angelevska et al., 2021; Ghasempour et al., 2021). Carbon monoxide (CO), nitrogen dioxide (NO 2), ozone (O 3), sulfur dioxide (SO 2), and particulate matter (including PM 10 and PM 2.5) emissions from human and natural sources have a substantial influence on individual health and well-being (Gopalakrishnan et al., 2018). With more than half of the world's population living in cities, the impact of air pollution on public health must be addressed. Energy consumption, industrial emissions, and automobile traffic all rise when cities develop in population and size, all of which can have a negative impact on air quality (Kahyaoğlu-Koračin et al., 2009). Conversion of forests, grasslands, and cropland to urban development, industrial complexes, and big commercial areas frequently results in increased emissions. Urban sprawl is the most severe example of this sort of growth, characterized by dispersed patterns of low-density development, which is frequently automobile-oriented (Superczynski & Christopher, 2011). Inevitably, air quality varies based on land cover and as the environment changes. The air quality impacts of different land covers have been researched in a number of studies. Thus, the air quality impacts of grasslands and shrublands have been investigated in the USA (Gopalakrishnan et al., 2018) for the primary purpose of estimating the pollution removal capacity of canopy cover on a national level. Spatial interpolation of air pollution measurements was done, and the land cover was considered (Janssen et al., 2008). As a land use Abstract With the increased urbanization, the rise of the manufacturing industry, and the use of fossil fuels, poor air quality is one of the most serious and pressing problems worldwide. The COVID-19 outbreak prompted absolute lockdowns in the majority of countries throughout the world, posing new research questions. The study's goals were to analyze air and temperature parameters in Turkey across various land cover classes and to investigate the correlation between air and temperature. For that purpose, remote sensing data from MODIS and Sentinel-5P TROPOMI were used from 2019 to 2021 over Turkey. A large amount of data was processed and analyzed in Google Earth Engine (GEE). Results showed a significant decrease in NO 2 in urban areas. The findings can be used in long-term strategies for lowering global air pollution. Future research should look at similar investigations in various study sites and evaluate changes in air metrics over additional classes.
International journal of environment and geoinformatics, Apr 12, 2019
Remote sensing technologies provide very important big data to various science areas such as risk... more Remote sensing technologies provide very important big data to various science areas such as risk identification, damage detection and prevention studies. However, the classification processes used to create thematic maps to interpret this data can be ineffective due to the wide range of properties that these images provide. At this point, there arises a requirement to optimize the data. The first objective of this study is to evaluate the performance of the Bat Search Algorithm which has not previously been used for improving the classification accuracy of remotely sensed images by optimizing attributes. The second objective is to compare the performance of the Genetic Algorithm, Bat Search Algorithm, Cuckoo Search Algorithm and Particle Swarm Optimization Algorithm, which are used in many areas of the literature for the optimization of the attributes of remotely sensed images. For these purposes, an image from the Landsat 8 satellite is used. The performance of the algorithms is compared by classifying the image using the K-Means method. The analysis shows a 10-22% increase in overall accuracy with the addition of attribute optimization.
Anadolu Üniversitesi bilim ve teknoloji dergisi -b- teorik bilimler, Dec 31, 2018
Binalar deprem afetinden en fazla etkilenen nesnelerdir. Deprem sonrası yıkılan binaların tespit ... more Binalar deprem afetinden en fazla etkilenen nesnelerdir. Deprem sonrası yıkılan binaların tespit edilmesi, hem mevcut durumunun belirlenmesi hem de hızlı müdahale açısından önemlidir. Son yıllarda gelişen insansız hava araçları, üzerlerine takılan kamera sistemleri sayesinde yeryüzüne ait çok yüksek çözünürlüklü görüntüler elde edilebilmektedir. Bu görüntülerden üretilen ürünler aracılığı ile istenilen amaca yönelik bilgiler çıkarılabilmektedir. Bu çalışmada, 2015 ve 2014 yıllarında insansız hava aracı ile yüksek çözünürlüklü görüntüleri elde edilen bir alanda, yıkılan binaların tespiti gerçekleştirilmiştir. Bina tespiti işlemi senaryo bir olay üzerinden yapılmıştır. Bu kapsamda, 2015 yılı görüntüleri deprem öncesi, 2014 yılı görüntüleri deprem sonrası olarak ele alınmıştır. Her iki yıla ait görüntüler işlenerek alana ait sayısal yükseklik modeli ve ortofoto görüntü üretilmiştir. Üretilen bu verilere nesne tabanlı sınıflandırma işlemi uygulanarak, çalışma alanında yer alan binalar çıkarılmıştır. Her iki yıla ait bina sınıflarının karşılaştırılması ile 2015 yılında alanda mevcut olup, 2014 yılında alanda olmayan 11 bina başarı ile tespit edilmiştir.
Forest fires cause a lot of damage in Turkey, due to the geographical location, high temperatures... more Forest fires cause a lot of damage in Turkey, due to the geographical location, high temperatures, and low humidity levels. Accurate determination of burned forest areas is crucial for correct damage assessment studies, fire risk calculations, and review of the forest regeneration processes. In this study, we compare the performances of unsupervised classification methods (which have not been used to map burned areas before) of burned area extraction from medium resolution satellite images with K-means. In this regard, the areas affected by fire in the Kumluca and Adrasan regions in 2016, Alanya and Gümüldür regions in 2017 and Athens region in 2018 are determined using Landsat 8 images. For this purpose, Canopy, M-tree, a hierarchical clustering algorithm, and a learning vector quantization which are frequently used in the literature are applied to determine the burned area, and the results obtained are compared with the results obtained from K-means. The results show that unsupervised classification methods can be used to map burned areas. The hierarchical clustering and K-means algorithms provide the most accurate results in mapping burned areas in most of the regions used in the study.
Değişim tespitinin temel amacı, aynı bölgenin farklı zamanlarda çekilmiş görüntülerini karşılaştı... more Değişim tespitinin temel amacı, aynı bölgenin farklı zamanlarda çekilmiş görüntülerini karşılaştırarak her piksele değişim olan ve olmayan olmak üzere ikili kodlanmış etiketler atamaktır. Yüksek çözünürlüklü optik uzaktan algılama görüntülerine dayalı değişim tespiti çalışma alanında bulunan nesnelerin karmaşıklığı ve iki tarih arasındaki farklı görüntüleme koşulları nedeniyle zorlu bir görevdir. Bu çalışmada LEarning, VIsion and Remote sensing (LEVIR)-CD veri seti ile eğitilmiş, derin sinir ağı temelli Bitemporal Görüntü Dönüştürücü (Bitemporal Image Transformer-BIT) ve STANet modellerinin farklı çalışma alanlarındaki performansının araştırılması amaçlanmıştır. Elde edilen sonuçlar LEVIR-CD veri seti ile eğitilmiş olan BIT ve STANet modellerinin yüksek doğruluk ile değişim tespiti gerçekleştirmesi için ek eğitim veri setine ihtiyaç duyduğunu göstermektedir.
Süper çözünürlük, çeşitli yollarla görüntü çözünürlüğünü artırmayı amaçlayan ve son yıllarda deri... more Süper çözünürlük, çeşitli yollarla görüntü çözünürlüğünü artırmayı amaçlayan ve son yıllarda derin öğrenme alanındaki gelişmelerle beraber daha iyi sonuçların elde edildiği bir görüntü iyileştirme yöntemidir. Keşif-gözetleme, nesne tespiti, çeşitli tıp uygulamaları gibi birçok alanda kendine uygulama sahası bulan bu yöntemle düşük çözünürlüklü görüntülerden daha fazla bilgi çıkarımı yapmak mümkün hale gelmektedir. Görüntü iyileştirme işlemi geleneksel yollarda interpolasyon gibi matematiksel tahminleme yöntemleriyle başarılmaya çalışılırken derin öğrenmeye dayalı modeller bunu çeşitli Konvolüsyonel Sinir Ağı mimarilerini etiketlenmiş verilerle beraber kullanarak gerçekleştirmektedir. Bu çalışmada yaklaşık 3 metre yersel çözünürlüğe sahip 8-bitlik PlanetScope uydu görüntüleri biri geleneksel bikübik interpolasyon, diğeri derin öğrenmeye dayalı ESPCN (Efficient Sub-Pixel Convolutional Neural Network) modelleri kullanılarak iyileştirilmiş ve sonuçlar PSNR (Peak Signal to Noise Ratio) değerleri cinsinden karşılaştırılmıştır. ESPCN'nin eğitiminde farklı veri setleri kullanılmış ve neticeye etkisi gözlenmiştir. Elde edilen sonuçlara göre; eğitimde veri sayısını artırmanın ve eğitim verisi türünün test verisiyle benzer olmasının sonuçları olumlu etkilediği ve ayrıca her ne kadar farklı türde ve az sayıda veriyle eğitilmiş olsa bile ESPCN yönteminin geleneksel bikübik interpolasyona göre daha iyi sonuçlar ürettiği ortaya konulmuştur.
Su kaynakları yaşamın devamlılığında önemli bir rol oynamaktadır. Bu nedenle su kaynaklarının har... more Su kaynakları yaşamın devamlılığında önemli bir rol oynamaktadır. Bu nedenle su kaynaklarının haritalanması ve değişikliklerin izlenmesi gerekmektedir. Su kaynaklarının izlenmesi, kontrolü ve koruma çalışmalarında uzaktan algılama teknolojileri önemli veriler sağlamaktadır. Bu veriler, su kütleleri ile ilgili çalışmalarda planlayıcılar için önemlidir. Bu çalışmada Manisa'ya 70 km uzaklıkta bulunan Gölmarmara ilçesinde yer alan Marmara Gölü su yüzeyinin değişim analizi gerçekleştirilmiştir. Ek olarak Marmara Gölünün gelecekteki alansal değişimine ait tahminleme çalışması gerçekleştirilmiştir. Bu doğrultuda yüzey alanları, çalışma alanına ait 2002-2021 yıllarına ait Landsat 7 görüntülerinin kontrolsüz sınıflandırma yöntemi ile analizi sonucunda elde edilmiştir. Bunun yanında alana ait yağış, sıcaklık ve arazi yüzey sıcaklığı (LST) verileri Google Earth Engine yardımıyla elde edilmiştir. Elde edilen veriler kullanılarak en doğru tahminlemeyi yapabilmek amacıyla Radyal Tabanlı Fonksiyon (RBF Regressor), Doğrusal Regresyon (Lineer Regression), Toplamsal Regresyon (Additive Regression) ve Çok Katmanlı Perceptron Sınıflandırıcı (MultiLayer Perceptron Classifier) yöntemleri kullanılmıştır. 2002-2012 yılları arasındaki veriler kullanılarak 2013 ve 2021 yılları arasındaki değişim belirlenmiştir. Sonuçlar incelendiğinde en iyi tahminin R2= 0.91 ile Çok Katmanlı Perceptron CS ile elde edildiği gözlemlenmiştir. Bu yöntem ile 2022 ve 2026 yılları için gerçekleştirilen tahmin çalışması sonucunda gölün çok daha fazla küçüleceği ve 1.56 km2' ye ulaşacağı öngörülmüştür.
This article is an open access article distributed under the terms and conditions of the Creative... more This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY
Computers, Environment and Urban Systems, Jul 1, 2018
Geocoding is a tool that can be used in many areas such as the development of disaster prevention... more Geocoding is a tool that can be used in many areas such as the development of disaster prevention systems, crime mapping and the monitoring of communicable diseases, and which has gradually gained importance. However, the use of geocoding is not yet possible in some areas where it could serve as an effective tool, for various reasons such as inconsistencies in address formats, including inaccurate numbering systems, misspellings, the use of abbreviations and a lack of data that refers to the geocoding process. This study seeks to address these problems by way of a standardization process. To that end, it employs a method that decomposes addresses used as input data in geocoding, identifies spelling mistakes and abbreviations, and reorganizes the addresses through the Natural Language Process (NLP). As test data, the addresses of primary schools in the district of Eskisehir are taken. First the geocoding process is performed on the data set, using both Google geocoding API and ArcGIS geocoding API. Then, the addresses are reformatted into three address formats by applying standardization processes. Geocoding is performed on the re-formatted addresses and the results compared to the non-standardized results. The standardization used is shown to make a significant improvement in the accuracy of the geocoding results. The method used in this study is significant not only in increasing the accuracy of the geocoding process, but also in sustaining its wider use.
Su kaynakları yaşamın devamlılığında önemli bir rol oynamaktadır. Bu nedenle su kaynaklarının har... more Su kaynakları yaşamın devamlılığında önemli bir rol oynamaktadır. Bu nedenle su kaynaklarının haritalanması ve değişikliklerin izlenmesi gerekmektedir. Su kaynaklarının izlenmesi, kontrolü ve koruma çalışmalarında uzaktan algılama teknolojileri önemli veriler sağlamaktadır. Bu veriler, su kütleleri ile ilgili çalışmalarda planlayıcılar için önemlidir. Bu çalışmada Manisa'ya 70 km uzaklıkta bulunan Gölmarmara ilçesinde yer alan Marmara Gölü su yüzeyinin değişim analizi gerçekleştirilmiştir. Ek olarak Marmara Gölünün gelecekteki alansal değişimine ait tahminleme çalışması gerçekleştirilmiştir. Bu doğrultuda yüzey alanları, çalışma alanına ait 2002-2021 yıllarına ait Landsat 7 görüntülerinin kontrolsüz sınıflandırma yöntemi ile analizi sonucunda elde edilmiştir. Bunun yanında alana ait yağış, sıcaklık ve arazi yüzey sıcaklığı (LST) verileri Google Earth Engine yardımıyla elde edilmiştir. Elde edilen veriler kullanılarak en doğru tahminlemeyi yapabilmek amacıyla Radyal Tabanlı Fonksiyon (RBF Regressor), Doğrusal Regresyon (Lineer Regression), Toplamsal Regresyon (Additive Regression) ve Çok Katmanlı Perceptron Sınıflandırıcı (MultiLayer Perceptron Classifier) yöntemleri kullanılmıştır. 2002-2012 yılları arasındaki veriler kullanılarak 2013 ve 2021 yılları arasındaki değişim belirlenmiştir. Sonuçlar incelendiğinde en iyi tahminin R2= 0.91 ile Çok Katmanlı Perceptron CS ile elde edildiği gözlemlenmiştir. Bu yöntem ile 2022 ve 2026 yılları için gerçekleştirilen tahmin çalışması sonucunda gölün çok daha fazla küçüleceği ve 1.56 km2' ye ulaşacağı öngörülmüştür.
International Journal of Engineering and Geosciences
Remote sensing Google Earth Engine Forest areas changes Statistical Analyze Forest area losses ar... more Remote sensing Google Earth Engine Forest areas changes Statistical Analyze Forest area losses are one of the most significant changes in land cover. These losses negatively affect ecosystems and cause severe economic and social life problems. It is necessary to monitor the process carefully and analyze the effects well to minimize all these negative effects in forest land losses and improve the development in urban areas positively. It is of great importance that these analyses are carried out quickly and accurately in terms of developing the natural environment. In this study, the effects that cause forest losses in the Mediterranean Region over the years are examined with the data obtained with the Google Earth Engine (GEE). Within the scope of the study, the changes in forest areas in the Mediterranean Region between 2004 and 2019 have been examined by considering many factors. In the study, Normalized Difference Vegetation Index (NDVI), precipitation, temperature, land surface temperature, aerosol optical depth, ozone, fire, urban areas, and population data were obtained with GEE. The data obtained were analyzed statistically, and the factors affecting the losses in forest areas the most were determined.
With the developing technology and automation, automatic labelling of images is of great importan... more With the developing technology and automation, automatic labelling of images is of great importance for automatic mapping. However, the most significant disadvantage of this method is that the classes’ labels cannot be generated automatically. In the current remote sensing literature, understanding and automatically labelling clusters obtained from the clustering process without a training phase is a problem that requires effective solutions. In this study, in order to solve this problem, we present a methodology that creates labels without any training phase. We use the bands in the image and Corine data in this process. The methodology uses a database created by examining the spectral characteristics of land classes from sample images collected from various geographies and time periods. The spectral index values of the unlabelled classes obtained are evaluated using this database, and the relevant label is assigned to each class. This database was created by analyzing Sentinel-2 Level-1 images of the Mediterranean and the Black Sea regions in Turkey. Then, these labels compare with the Corine classes corresponds to each pixel according to the ruleset. This developed approach aims to automatically label land, a green agricultural area, forest, urban area, and uncultivated agricultural area. The reason for choosing these areas is that they are the areas that generally make up the environment and a large part of the ecosystem, which are important areas that many researchers frequently use in their studies. The methodology developed was tested with Sentinel 2 images of Gemlik, Hatay regions from Turkey, and Agioi Apostoli region from Greece. The results of the accuracy analysis are 80%, 83%, and %82 for Gemlik, Hatay, and Agioi Apostoli areas.
Journal of the Indian Society of Remote Sensing, 2022
Urban areas of major cities in developing countries are expanding rapidly due to rapid population... more Urban areas of major cities in developing countries are expanding rapidly due to rapid population growth and industrialization. Determining the reasons for rapid urbanization and the amount of urban area that will be needed in the future is important for the planned growth of these cities. In this study, the LU/LC change of Eskisehir city center that has a rapid rate of urbanization and industrialization in Turkey between 1984 and 2020 was determined. Then, the area needed for urban and industrial areas in 2030 was investigated using the Cullullar Automata–Markov Chain (CA–MC) hybrid model simulation. Each year's LU/LC map was produced with over 80% kappa accuracies performing the pixel-based Random Forest (RF) algorithm for change analysis. In the change analysis made between 1984 and 2020, it was determined that there was a change of 117% in the urban area and 977% in the industrial area. The prediction was made for the validation of the CA–MC model from the period 2000–2010 to 2020. When compared with the 2020 LU/LC map, the model success was obtained as 0.84, 0.87, and 0.87 in the Kstandard, Klocation, and Kno Kappa metrics, respectively. By using the 2010–2020 periods in the estimation of 2030, it has been observed that the urban and industrial area will increase by 22.95%; therefore, there will be decreases in agriculture and other natural areas as in previous years.
Monitoring trees is necessary to manage and take inventory of forests, monitor plants in urban ar... more Monitoring trees is necessary to manage and take inventory of forests, monitor plants in urban areas, distribute vegetation, monitor change, and establish sensitive and renewable agricultural systems. This study aims to automatically detect, count, and map apricot trees in an orthophoto, covering an area of approximately 48 ha on the ground surface using two different algorithms based on deep learning. Here, Mask region-based convolutional neural network (Mask R-CNN) and U-Net models were run together with a dilation operator to detect apricot trees in UAV images, and the performances of the models were compared. Results show that Mask R-CNN operated in this way performs better in tree detection, counting, and mapping tasks compared to U-Net. Mask R-CNN with the dilation operator achieved a precision of 98.7%, recall of 99.7%, F1 score of 99.1%, and intersection over union (IoU) of 74.8% for the test orthophoto. U-Net, on the other hand, has achieved a recall of 93.3%, precision of ...
Uploads
Papers by dilek matci