Cengiz Avci
Cengiz Avci has a total of 5 co-authors. Here are the top co-authors by number of shared publications:
- Elif Sertel
Istanbul Technical University
1 shared publication - Filiz Bektas Balcik
Istanbul Technical University
1 shared publication - Muhammed Budak
ISTANBUL TECHNICAL UNIVERSITY, FACULTY OF CIVIL ENGINEERING, DEPARTMENT OF GEOMATICS ENGINEERING, GEOMATICS ENGINEERING PR
1 shared publication - Mustafa Erdem Kabadayi
Koc University
1 shared publication - Nur Yagmur
Engineering Faculty, Geomatics Engineering Department, Gebze Technical University, Kocaeli, Türkiye
1 shared publication
Latest Publications
- Top Cited Papers
- 15 February 2023
- Published by International Journal of Engineering and Geoscience in International Journal of Engineering and Geosciences
- Vol. 8 (1) , 1-10
Comparison between random forest and support vector machine algorithms for LULC classification
Nowadays, machine learning (ML) algorithms have been widely chosen for classifying satellite images for mapping Earth's surface. Support Vector Machine (SVM) and Random Forest (RF) stand out among these algorithms with their accurate results in the literature. The aim of this study is to analyze the performances of these algorithms on land use and land cover (LULC) classification, especially wetlands which have significant ecological functions. For this purpose, Sentinel-2 satellite image, which is freely provided by European Space Agency (ESA), was used to monitor not only the open surface water body but also around Marmara Lake. The performance evaluation was made with the increasing number of the training dataset. 3 different training datasets having 10, 15, and 20 areas of interest (AOI) per class, respectively were used for the classification of the satellite images acquired in 2015 and 2020. The most accurate results were obtained from the classification with RF algorithm and 20 AOIs. According to obtained results, the change detection analysis of Marmara Lake was investigated for possible reasons. Whereas the water body and wetland have decreased more than 50% between 2015 and 2020, crop sites have increased approximately 50%.
- 15 February 2023
- Published by International Journal of Engineering and Geoscience in International Journal of Engineering and Geosciences
- Vol. 8 (1) , 1-10
Top Cited PapersComparison between random forest and support vector machine algorithms for LULC classification
Nowadays, machine learning (ML) algorithms have been widely chosen for classifying satellite images for mapping Earth's surface. Support Vector Machine (SVM) and Random Forest (RF) stand out among these algorithms with their accurate results in the literature. The aim of this study is to analyze the performances of these algorithms on land use and land cover (LULC) classification, especially wetlands which have significant ecological functions. For this purpose, Sentinel-2 satellite image, which is freely provided by European Space Agency (ESA), was used to monitor not only the open surface water body but also around Marmara Lake. The performance evaluation was made with the increasing number of the training dataset. 3 different training datasets having 10, 15, and 20 areas of interest (AOI) per class, respectively were used for the classification of the satellite images acquired in 2015 and 2020. The most accurate results were obtained from the classification with RF algorithm and 20 AOIs. According to obtained results, the change detection analysis of Marmara Lake was investigated for possible reasons. Whereas the water body and wetland have decreased more than 50% between 2015 and 2020, crop sites have increased approximately 50%.
- 8 September 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 19 (1545598X) , 1-5
Deep Learning-Based Road Extraction From Historical Maps
Automatic road extraction from historical maps is an important task to understand past transportation conditions and conduct spatiotemporal analysis revealing information about historical events and human activities over the years. This research aimed to propose the ideal architecture, encoder, and hyperparameter settings for the historical road extraction task. We used a dataset including 7076 patches with the size of $256 \times256$ pixels generated from scanned historical Deutsche Heereskarte 1:200 000 Türkei (DHK 200 Turkey) maps and their corresponding digitized ground truth masks for five different roads types. We first tested the widely used Unet++ and Deeplabv3 architectures. We also evaluated the contribution of attention models by implementing Unet++ with the concurrent spatial and channel-squeeze and excitation block and multiscale attention net. We achieved the best results with split-attention network (Timm-resnest200e) encoder and Unet++ architecture, with 98.99% overall accuracy, 41.99% intersection of union, 51.41% precision, 69.7% recall, and 57.72% F1 score values. Our output weights could be directly used for the inference of other DHK maps and transfer learning for similar or different historical maps. The proposed architecture could also be implemented in different road extraction studies.
- 8 September 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 19 (1545598X) , 1-5
Deep Learning-Based Road Extraction From Historical Maps
Automatic road extraction from historical maps is an important task to understand past transportation conditions and conduct spatiotemporal analysis revealing information about historical events and human activities over the years. This research aimed to propose the ideal architecture, encoder, and hyperparameter settings for the historical road extraction task. We used a dataset including 7076 patches with the size of $256 \times256$ pixels generated from scanned historical Deutsche Heereskarte 1:200 000 Türkei (DHK 200 Turkey) maps and their corresponding digitized ground truth masks for five different roads types. We first tested the widely used Unet++ and Deeplabv3 architectures. We also evaluated the contribution of attention models by implementing Unet++ with the concurrent spatial and channel-squeeze and excitation block and multiscale attention net. We achieved the best results with split-attention network (Timm-resnest200e) encoder and Unet++ architecture, with 98.99% overall accuracy, 41.99% intersection of union, 51.41% precision, 69.7% recall, and 57.72% F1 score values. Our output weights could be directly used for the inference of other DHK maps and transfer learning for similar or different historical maps. The proposed architecture could also be implemented in different road extraction studies.