Akademik Veri Yönetim Sistemi
19th International Conference on Geoinformatics: Theoretical and Applied Aspects, Geoinformatics 2020, Kiev, Ukrayna, 11 - 14 Mayıs 2020
© Geoinformatics 2020 - XIXth International Conference "Geoinformatics: Theoretical and Applied Aspects". All rights reserved.Ground Penetrating Radar (GPR) is a geophysical technique that investigates underground structures. GPR data are used to visualize an underground map of search areas. One of the most important problems in the investigation of the underground structure is the correct identification, visualization and interpretation of anomalies in GPR data. To contribute to the solution of this problem, a methodology which consists of four stages is proposed. At the first stage, the original data from the test area are collected and preprocessed, and the synthetic data are produced using original data. In the second stage, features are extracted and the data set is created to apply the machine learning techniques. In the third stage, the data set is analyzed by machine learning techniques to identify the anomalies. At the final stage, the geometry of the anomalies in GPR data is visualized in the 3D environment and visualized anomalies are interpreted. The obtained results showed that the anomalies were detected with 94% accuracy.