Akar, Özlem; Saralioğlu, Ekrem; Güngör, Oğuz; Bayata, Halim Ferit

The Erzincan (Cimin) grape, which is an endemic product, plays a significant role in the economy of both the region it is cultivated in and the overall country. Therefore, it is crucial to closely monitor and promote this product. The objective of this study was to analyze the spatial distribution of vineyards by utilizing advanced machine learning and deep learning algorithms to classify high-resolution satellite images. A deep learning model based on a 3D Convolutional Neural Network (CNN) was developed for vineyard classification. The proposed model was compared with traditional machine learning algorithms, specifically Support Vector Machine (SVM), Random Forest (RF), and Rotation Forest (ROTF). The accuracy of the classifications was assessed through error matrices, kappa analysis, and McNemar tests. The best overall classification accuracies and kappa values were achieved by the 3D CNN and RF methods, with scores of 86.47% (0.8308) and 70.53% (0.6279) respectively. Notably, when Gabor texture features were incorporated, the accuracy of the RF method increased to 75.94% (0.6364). Nevertheless, the 3D CNN classifier outperformed all others, yielding the highest classification accuracy with an 11% advantage (86.47%). The statistical analysis using McNemar's test confirmed that the χ2 values for all classification outcomes exceeded 3.84 at the 95% confidence interval, indicating a significant enhancement in classification accuracy provided by the 3D CNN classifier. Additionally, the 3D CNN method demonstrated successful classification performance, as evidenced by the minimum-maximum F1-score (0.79-0.97), specificity (0.95-0.99), and accuracy (0.91-0.99) values.

Akar, A.; Gokalp, E.

The purpose of this study is to identify the deficiencies of the rangeland information system currently used in Turkey and, as an alternative, design a sustainable rangeland information system. In the study, both the extent of changes that occurred over time in the rangelands and the factors that caused such changes were identified, and solutions were suggested to eliminate those factors. The rangelands located in the Akcaabat district of Trabzon province were selected as the study area. Land use maps were produced by using the object-based classification method. According to the results of change analyses made with this information system, it was found out that, from 1973 to 2012, a surface area of 159.8 hectares had been degraded, demonstrating that the current information system had not been successful enough in the management of rangelands. For that reason, a sustainable rangeland information system free from all deficiencies was designed.

Akar, A.

The main objective of the study was to examine accuracies of DEMs (Digital Elevation Models) with different topographical structures generated by using the Unmanned Aerial Vehicle (UAV) point clouds. Two different terrains with flat and sloping topographical structures were selected for the study, and DEMs of these terrains were generated using eight interpolation techniques (Kriging, Natural Neighbor, Radial Basis Function Triangulation with Linear interpolation, Nearest Neighbor, Invers Distance to a Power, Local Polynomial and Minimum Curvature). The accuracies of DEMs were tested by calculating the statistic methods with the help of the control points obtained by land surveying techniques. At the end of the study, it was observed that in DEMs prepared for both flat (study area 1) and sloping (study area 2) terrains, Kriging interpolation method yields the best results as study area 1 and 2, respectively. In addition, the results were examined using Shapiro-Wilk and ANOVA: Friedman tests. After observing with the Shapiro-Wilk test that the data has a normal distribution, it was statistically determined through the parametric ANOVA: Friedman test that there is no difference between the variables.