Nafiye Nur Apaydın
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Nafiye Nur Apaydın has a total of 13 co-authors. Here are the top co-authors by number of shared publications:
- Orhan Yaman
Department of Digital Forensics Engineering, Firat University, 23119 Elazig, Turkey
7 shared publications - İrFan Kilic
Firat Universitesi
4 shared publications - Mehmet Karakose
Computer Engineering Department, Firat University, 23119 Elazığ, Türkiye
3 shared publications - Abdulsamed Güneş
Firat Universitesi
1 shared publication - Adnan Topuz
Department of Mechanical Engineering, Faculty of Engineering, Zonguldak Bülent Ecevit University, 23119 Elazig, Turkey
1 shared publication - Beytullah Erdoğan
Department of Mechanical Engineering, Zonguldak Bulent Ecevit University, Zonguldak 67100, Turkey
1 shared publication
Latest Publications
- Open Access
- 20 September 2026
- Published by MDPI AG in Sustainability
- Vol. 18 (18) , 9641
A Hybrid Multispectral Deep Feature Fusion Framework for Sustainable Solid Waste Classification Using RGB and NIR Imaging
The rapid increase in solid waste resulting from rising consumption has increased the need for effective recycling technologies to reduce environmental impacts and recover valuable resources. This study proposes a hybrid approach based on RGB and near-infrared (NIR) imaging for the automated classification of solid waste. A unique dataset consisting of five waste classes—glass, paper, fabric, metal, and plastic—was created using a conveyor belt system equipped with synchronized RGB and NIR cameras. Waste objects were detected in RGB images using the Segment Anything Model (SAM) and cropped according to the generated YOLO labels. Three spectral channel combinations, NIR-G-B, R-NIR-B, and R-G-NIR, were generated from paired RGB and NIR images. Deep features were extracted from RGB, NIR, and these spectral combinations using DenseNet121. Feature dimensionality was then reduced using PCA, and the resulting features were classified using SVM. The proposed method achieved 98.01% accuracy and recall on the 80/20 hold-out test split, outperforming models based solely on RGB or NIR images, while 5-fold and 10-fold cross-validation yielded mean accuracies of 96.16% and 96.34%, respectively. The results demonstrate that combining the complementary features of visible and NIR spectral information is effective in distinguishing different types of waste. The proposed approach can contribute to the development of smart and sustainable recycling systems by supporting more accurate and automated sorting of recyclable materials.
- 20 September 2026
- Published by MDPI AG in Sustainability
- Vol. 18 (18) , 9641
Open AccessA Hybrid Multispectral Deep Feature Fusion Framework for Sustainable Solid Waste Classification Using RGB and NIR Imaging
The rapid increase in solid waste resulting from rising consumption has increased the need for effective recycling technologies to reduce environmental impacts and recover valuable resources. This study proposes a hybrid approach based on RGB and near-infrared (NIR) imaging for the automated classification of solid waste. A unique dataset consisting of five waste classes—glass, paper, fabric, metal, and plastic—was created using a conveyor belt system equipped with synchronized RGB and NIR cameras. Waste objects were detected in RGB images using the Segment Anything Model (SAM) and cropped according to the generated YOLO labels. Three spectral channel combinations, NIR-G-B, R-NIR-B, and R-G-NIR, were generated from paired RGB and NIR images. Deep features were extracted from RGB, NIR, and these spectral combinations using DenseNet121. Feature dimensionality was then reduced using PCA, and the resulting features were classified using SVM. The proposed method achieved 98.01% accuracy and recall on the 80/20 hold-out test split, outperforming models based solely on RGB or NIR images, while 5-fold and 10-fold cross-validation yielded mean accuracies of 96.16% and 96.34%, respectively. The results demonstrate that combining the complementary features of visible and NIR spectral information is effective in distinguishing different types of waste. The proposed approach can contribute to the development of smart and sustainable recycling systems by supporting more accurate and automated sorting of recyclable materials.
- 29 June 2026
- Published by Journal of Physical Chemistry and Functional Materials in Journal of Physical Chemistry and Functional Materials
- Vol. 9 (1) , 87-95
Vehicle Color-Type Detection Based on YOLOv8 and K-Means
There is provided a computing device for managing honeybee colonies for pollination of crop(s), comprises: a processor(s) configured for: for each one of a plurality of honeybee colonies positioned for pollination of crop(s) in a geographical area: obtaining in-colony feature(s), computed based on output of internal sensor(s) monitoring the honeybee colony, wherein the in-colony feature(s) is indicative of an internal state of the honeybee colony positioned for pollination of crop(s) in the geographical area, obtaining out-colony feature(s), computed based on output of external sensor(s) monitoring the environment of the honeybee colony, wherein the out-colony feature is indicative of an external environment of the honeybee colony, feeding a combination of the in-colony feature(s) and the out-colony feature(s) into a machine learning model, and obtaining an outcome indicating pollination effectiveness of the honeybee colony.
- 29 June 2026
- Published by Journal of Physical Chemistry and Functional Materials in Journal of Physical Chemistry and Functional Materials
- Vol. 9 (1) , 87-95
Vehicle Color-Type Detection Based on YOLOv8 and K-Means
There is provided a computing device for managing honeybee colonies for pollination of crop(s), comprises: a processor(s) configured for: for each one of a plurality of honeybee colonies positioned for pollination of crop(s) in a geographical area: obtaining in-colony feature(s), computed based on output of internal sensor(s) monitoring the honeybee colony, wherein the in-colony feature(s) is indicative of an internal state of the honeybee colony positioned for pollination of crop(s) in the geographical area, obtaining out-colony feature(s), computed based on output of external sensor(s) monitoring the environment of the honeybee colony, wherein the out-colony feature is indicative of an external environment of the honeybee colony, feeding a combination of the in-colony feature(s) and the out-colony feature(s) into a machine learning model, and obtaining an outcome indicating pollination effectiveness of the honeybee colony.
- Open Access
- 21 January 2026
- Published by MDPI AG in Sustainability
- Vol. 18 (2) , 1067
A Sustainability-Focused Real-Time Dynamic Wind Speed Estimation Method for Turbine Performance Optimization
To achieve the highest efficiency from the turbines used in wind power plants, the region where the plant will be located must meet the appropriate conditions. One of these conditions, and the most important, is that the wind potential be above the critical value for energy production and be continuous. Locations that meet these conditions contribute positively to energy production and produce high efficiency. Based on the interpreted data, temperature, wind direction, and wind speed data from three turbines located at altitudes of 432, 454, and 492 m in the Sebenoba area of Yayladağ, Hatay, where wind potential is high, were collected at 10 min intervals between 1 January 2017, and 19 September 2018, yielding a total of 50,986 data points. Wind speed was estimated for this region using temperature, wind direction, and time information. Daily, monthly, and seasonal analyses were used to generate forecasts for the three altitudes. Wind speed was estimated using Decision Tree Regression and 10-Fold Cross Validation methods, and Root Mean Square Error (RMSE) values were found to be 0.64917, 0.66629, and 0.59954 for the three altitudes, respectively; the overall RMSE value was found to be 0.60188. RMSE values decreased in daily, monthly, and seasonal analyses, and an inverse relationship existed between wind speed and RMSE. Analysis of these results indicated that the forecast model was suitable. This study supports sustainability by enabling accurate wind speed forecasting for optimal turbine placement, improving energy efficiency, and promoting long-term environmentally and economically sustainable wind energy planning.
- 21 January 2026
- Published by MDPI AG in Sustainability
- Vol. 18 (2) , 1067
Open AccessA Sustainability-Focused Real-Time Dynamic Wind Speed Estimation Method for Turbine Performance Optimization
To achieve the highest efficiency from the turbines used in wind power plants, the region where the plant will be located must meet the appropriate conditions. One of these conditions, and the most important, is that the wind potential be above the critical value for energy production and be continuous. Locations that meet these conditions contribute positively to energy production and produce high efficiency. Based on the interpreted data, temperature, wind direction, and wind speed data from three turbines located at altitudes of 432, 454, and 492 m in the Sebenoba area of Yayladağ, Hatay, where wind potential is high, were collected at 10 min intervals between 1 January 2017, and 19 September 2018, yielding a total of 50,986 data points. Wind speed was estimated for this region using temperature, wind direction, and time information. Daily, monthly, and seasonal analyses were used to generate forecasts for the three altitudes. Wind speed was estimated using Decision Tree Regression and 10-Fold Cross Validation methods, and Root Mean Square Error (RMSE) values were found to be 0.64917, 0.66629, and 0.59954 for the three altitudes, respectively; the overall RMSE value was found to be 0.60188. RMSE values decreased in daily, monthly, and seasonal analyses, and an inverse relationship existed between wind speed and RMSE. Analysis of these results indicated that the forecast model was suitable. This study supports sustainability by enabling accurate wind speed forecasting for optimal turbine placement, improving energy efficiency, and promoting long-term environmentally and economically sustainable wind energy planning.
- 1 December 2025
- Published by Igdir University in Iğdır Üniversitesi Fen Bilimleri Enstitüsü Dergisi
- Vol. 15 (4) , 1165-1177
Akan Görüntülerde Kavga tespiti için DenseNet121 ve Transfer Öğrenme Tabanlı Video Analiz Yöntemi
Ani gelişen kavga olaylarının erken tespiti ve bu olaylara hızlı ve etkili bir şekilde müdahale edilmesi, kamu güvenliğinin sağlanması ve olası olumsuz durumların önlenmesi açısından büyük bir öneme sahiptir. Günümüz teknolojik imkanları doğrultusunda, video tabanlı otomatik kavga tespit sistemleri, güvenlik güçlerine ve yetkililere zamanında uyarılar sağlayarak proaktif önlemler alınmasını mümkün kılmaktadır. Bu çalışmada, kavga (fight) ve kavga içermeyen (noFight) durumların doğru bir şekilde sınıflandırılabilmesi amacıyla derin öğrenme tabanlı bir yöntem önerilmektedir. Özellikle, görüntü işleme alanında başarılı sonuçlar elde eden DenseNet121 modeli kullanılarak transfer öğrenme yöntemiyle video tabanlı kavga tespiti gerçekleştirilmiştir. Deneysel çalışmalarda, literatürde yaygın olarak kullanılan Peliculas ve Hockey veri setleri değerlendirilmiş olup, önerilen yöntem ile Peliculas veri seti üzerinde %100 doğruluk (Accuracy), Hockey veri seti üzerinde ise %99.32 doğruluk elde edilmiştir. Elde edilen yüksek başarı oranları, önerilen yöntemin kavga tespitinde etkin bir şekilde kullanılabileceğini göstermektedir.
- 1 December 2025
- Published by Igdir University in Iğdır Üniversitesi Fen Bilimleri Enstitüsü Dergisi
- Vol. 15 (4) , 1165-1177
Akan Görüntülerde Kavga tespiti için DenseNet121 ve Transfer Öğrenme Tabanlı Video Analiz Yöntemi
Ani gelişen kavga olaylarının erken tespiti ve bu olaylara hızlı ve etkili bir şekilde müdahale edilmesi, kamu güvenliğinin sağlanması ve olası olumsuz durumların önlenmesi açısından büyük bir öneme sahiptir. Günümüz teknolojik imkanları doğrultusunda, video tabanlı otomatik kavga tespit sistemleri, güvenlik güçlerine ve yetkililere zamanında uyarılar sağlayarak proaktif önlemler alınmasını mümkün kılmaktadır. Bu çalışmada, kavga (fight) ve kavga içermeyen (noFight) durumların doğru bir şekilde sınıflandırılabilmesi amacıyla derin öğrenme tabanlı bir yöntem önerilmektedir. Özellikle, görüntü işleme alanında başarılı sonuçlar elde eden DenseNet121 modeli kullanılarak transfer öğrenme yöntemiyle video tabanlı kavga tespiti gerçekleştirilmiştir. Deneysel çalışmalarda, literatürde yaygın olarak kullanılan Peliculas ve Hockey veri setleri değerlendirilmiş olup, önerilen yöntem ile Peliculas veri seti üzerinde %100 doğruluk (Accuracy), Hockey veri seti üzerinde ise %99.32 doğruluk elde edilmiştir. Elde edilen yüksek başarı oranları, önerilen yöntemin kavga tespitinde etkin bir şekilde kullanılabileceğini göstermektedir.
- 30 October 2025
- Published by Duzce Universitesi Bilim ve Teknoloji Dergisi in Düzce Üniversitesi Bilim ve Teknoloji Dergisi
- Vol. 13 (4) , 1676-1694
Hybrid Color Space and GLCM Feature Extraction Based Classification Method for Underwater Image Analysis
In cases where man-made fishing nets are forgotten or lost underwater, these nets are called ghost nets. These ghost nets threaten the underwater ecosystem over time and reduce the biodiversity of living creatures underwater. For this reason, studies are being carried out to detect ghost nets underwater. In this study, SODD and Trash Icra data sets were used to detect net, trash, ROV and biological creature classes underwater. For each class, features were extracted using the GLCM filter and HSV, YUV, LAB and RGB color spaces. The extracted features were trained with the Random Forest Classification Algorithm and the results were obtained. As a result of the training, it was seen that although each color space had a low accuracy value on its own, when used together, it affected the performance positively and increased the accuracy, and the best accuracy value was 89.16% in the proposed method (HSV + YUV + LAB + RGB + GLCM). In addition, for the best case where all color spaces were used, Naive Bayes(NB), KNN and SVM classification algorithms were applied and the results were compared with the proposed method, Random Forest Classification. Accuracy values of 52% with NB, 62.17% with KNN and 73.67% with SVM were obtained and it was proven that the best method was the proposed method, Random Forest Classification algorithm. The results of the study demonstrate the effectiveness of integrating multi-color space features with texture analysis for underwater object classification, offering a promising approach for ghost net detection in real-world scenarios. Ghost nets, which are abandoned or lost man-made fishing nets, pose a significant threat to the marine ecosystem by entangling and endangering underwater life.
- 30 October 2025
- Published by Duzce Universitesi Bilim ve Teknoloji Dergisi in Düzce Üniversitesi Bilim ve Teknoloji Dergisi
- Vol. 13 (4) , 1676-1694
Hybrid Color Space and GLCM Feature Extraction Based Classification Method for Underwater Image Analysis
In cases where man-made fishing nets are forgotten or lost underwater, these nets are called ghost nets. These ghost nets threaten the underwater ecosystem over time and reduce the biodiversity of living creatures underwater. For this reason, studies are being carried out to detect ghost nets underwater. In this study, SODD and Trash Icra data sets were used to detect net, trash, ROV and biological creature classes underwater. For each class, features were extracted using the GLCM filter and HSV, YUV, LAB and RGB color spaces. The extracted features were trained with the Random Forest Classification Algorithm and the results were obtained. As a result of the training, it was seen that although each color space had a low accuracy value on its own, when used together, it affected the performance positively and increased the accuracy, and the best accuracy value was 89.16% in the proposed method (HSV + YUV + LAB + RGB + GLCM). In addition, for the best case where all color spaces were used, Naive Bayes(NB), KNN and SVM classification algorithms were applied and the results were compared with the proposed method, Random Forest Classification. Accuracy values of 52% with NB, 62.17% with KNN and 73.67% with SVM were obtained and it was proven that the best method was the proposed method, Random Forest Classification algorithm. The results of the study demonstrate the effectiveness of integrating multi-color space features with texture analysis for underwater object classification, offering a promising approach for ghost net detection in real-world scenarios. Ghost nets, which are abandoned or lost man-made fishing nets, pose a significant threat to the marine ecosystem by entangling and endangering underwater life.
- Open Access
- 30 April 2025
- Published by Acadlore Publishing Services Limited in Information Dynamics and Applications
- Vol. 4 (2) , 66-75
Crowd Density Estimation via a VGG-16-Based CSRNet Model
Accurate crowd density estimation has become critical in applications ranging from intelligent urban planning and public safety monitoring to marketing analytics and emergency response. In recent developments, various methods have been used to enhance the precision of crowd analysis systems. In this study, a Convolutional Neural Network (CNN)-based approach was presented for crowd density detection, wherein the Congested Scene Recognition Network (CSRNet) architecture was employed with a Visual Geometry Group (VGG)-16 backbone. This method was applied to two benchmark datasets—Mall and Crowd-UIT—to assess its effectiveness in real-world crowd scenarios. Density maps were generated to visualize spatial distributions, and performance was quantitatively evaluated using Mean Squared Error (MSE) and Mean Absolute Error (MAE) metrics. For the Mall dataset, the model achieved an MSE of 0.08 and an MAE of 0.10, while for the Crowd-UIT dataset, an MSE of 0.05 and an MAE of 0.15 were obtained. These results suggest that the proposed VGG-16-based CSRNet model yields high accuracy in crowd estimation tasks across varied environments and crowd densities. Additionally, the model demonstrates robustness in generalizing across different dataset characteristics, indicating its potential applicability in both surveillance systems and public space management. The outcomes of this investigation offer a promising direction for future research in data-driven crowd analysis, particularly in enhancing predictive reliability and real-time deployment capabilities of deep learning models for population monitoring tasks.
- 30 April 2025
- Published by Acadlore Publishing Services Limited in Information Dynamics and Applications
- Vol. 4 (2) , 66-75
Open AccessCrowd Density Estimation via a VGG-16-Based CSRNet Model
Accurate crowd density estimation has become critical in applications ranging from intelligent urban planning and public safety monitoring to marketing analytics and emergency response. In recent developments, various methods have been used to enhance the precision of crowd analysis systems. In this study, a Convolutional Neural Network (CNN)-based approach was presented for crowd density detection, wherein the Congested Scene Recognition Network (CSRNet) architecture was employed with a Visual Geometry Group (VGG)-16 backbone. This method was applied to two benchmark datasets—Mall and Crowd-UIT—to assess its effectiveness in real-world crowd scenarios. Density maps were generated to visualize spatial distributions, and performance was quantitatively evaluated using Mean Squared Error (MSE) and Mean Absolute Error (MAE) metrics. For the Mall dataset, the model achieved an MSE of 0.08 and an MAE of 0.10, while for the Crowd-UIT dataset, an MSE of 0.05 and an MAE of 0.15 were obtained. These results suggest that the proposed VGG-16-based CSRNet model yields high accuracy in crowd estimation tasks across varied environments and crowd densities. Additionally, the model demonstrates robustness in generalizing across different dataset characteristics, indicating its potential applicability in both surveillance systems and public space management. The outcomes of this investigation offer a promising direction for future research in data-driven crowd analysis, particularly in enhancing predictive reliability and real-time deployment capabilities of deep learning models for population monitoring tasks.
- Open Access
- 25 June 2024
- Published by Batman Universitesi in Batman Üniversitesi Yaşam Bilimleri Dergisi
- Vol. 14 (1) , 57-68
Decision Tree-Based Direction Detection Using IMU Data in Autonomous Robots
Location detection plays a crucial role in various applications. In this study, a machine learning (ML) method using inertial measurement unit (IMU) data was developed to determine direction with the Global Positioning System (GPS). In this study, an electronic board was designed using an Arduino Mega, Altimu-10 IMU sensor, GPS module, and SD card module. This electronic board was placed on a car to create a new dataset. This dataset consists of 1952x11 data. The dataset was obtained using accelerometer (x, y, z), gyroscope (x, y, z), compass (x, y, z), and GPS sensor data. The Decision Tree Algorithm was proposed for direction determination in this study. The angles between each position and the previous position were calculated using the latitude and longitude values obtained from the collected data. Then, the data were divided into 4 classes: North, East, South, and West, based on specific angle ranges. Finally, a direction detection model was developed using IMU data in the proposed method, achieving an accuracy of approximately 82.11%.
- 25 June 2024
- Published by Batman Universitesi in Batman Üniversitesi Yaşam Bilimleri Dergisi
- Vol. 14 (1) , 57-68
Open AccessDecision Tree-Based Direction Detection Using IMU Data in Autonomous Robots
Location detection plays a crucial role in various applications. In this study, a machine learning (ML) method using inertial measurement unit (IMU) data was developed to determine direction with the Global Positioning System (GPS). In this study, an electronic board was designed using an Arduino Mega, Altimu-10 IMU sensor, GPS module, and SD card module. This electronic board was placed on a car to create a new dataset. This dataset consists of 1952x11 data. The dataset was obtained using accelerometer (x, y, z), gyroscope (x, y, z), compass (x, y, z), and GPS sensor data. The Decision Tree Algorithm was proposed for direction determination in this study. The angles between each position and the previous position were calculated using the latitude and longitude values obtained from the collected data. Then, the data were divided into 4 classes: North, East, South, and West, based on specific angle ranges. Finally, a direction detection model was developed using IMU data in the proposed method, achieving an accuracy of approximately 82.11%.