|
|
- Investigation of the effect of satellite combinations on velocity
estimation by using Trimble RTX service
Authors: Deniz Öz Demir
Abstract: The earth's crust is in motion means that GNSS stations coordinates are also change. For this reason, the coordinates of GNSS stations and accuracies of the parameters which calculated based on these coordinates become important. In this study, online GNSS processing service Trimble RTX was used for two stations belongs to the MGEX (Multi-GNSS Experiment) network set up by IGS (International GNSS Service). 24-hour data is divided into 4, 6, 8 and 12 hour epochs and processed for GPS (G), GPS+GLONASS (GR) and GPS+GLONASS+Galileo (GRE) satellite combinations. Velocity components are estimated according to the least square estimation method using the coordinates obtained from 5-year of GNSS data by using different satellite combinations and different session durations. It has been observed that each of the combinations are consistent results in itself. However, when the results obtained from GPS data are accepted as true values and these are compared to other combinations, it can be said that the accuracy increased with not only GLONASS but also GLONASS and Galileo together.
PubDate: Fri, 03 May 2024 00:00:00 +030
- A new method proposal for calculating the easement price in high-voltage
transmission line expropriations
Authors: Seda Nur Marabaoğlu; Bayram Uzun
Abstract: Easement expropriation is carried out for high voltage transmission lines (HVTL) passing over real estates. This practice is guided by the Supreme Court decisions due to the inadequacy of the Expropriation Law No. 2942. A new decision is made for each application. Among these decisions, searching for an appropriate provision to address a newly encountered situation results in both labor and time loss. HVTL factors that negatively affect the value of real estate and the degree to which these factors affect the value are not clear and certain. This uncertainty causes easement price to be underestimated or overestimated. In the study, the factors affecting the real estate value of HVTL were determined as a result of interviews with local and foreign real estate appraisers, national and international surveys (with participants from 20 countries), review of Supreme Court decisions and expert reports, and evaluations with judges and lawyers. These factors were weighted using the Analytic Hierarchy Process. Thus, a HVTL impact-value model was developed that automatically calculates the easement price. The interface software for this automation model was created using Node.js, Electron, TypeScript, React, Ant Design and Webpack. The model was tested with 60 expert reports collected from various provinces in Turkey. 10 of these reports are shown as examples in the study. According to the automation model, the error rate in expert reports was found to be at least 17.21% and at most 84.56%. This model will provide a standard in calculating the easement price and will significantly eliminate the problem of under- or over-determination of this price. Since there is no defined workflow chart that users can follow step by step in determining the easement price, confusion and loss of time occur in the legal process. In order to eliminate this deficiency, evaluations were made with judges and lawyers and a workflow chart that would speed up the legal procedure in easement expropriation applications was created and presented in the study.
PubDate: Fri, 03 May 2024 00:00:00 +030
- Examining the accuracy of DEM of difference and 3D point cloud comparison
methods: Open pit mine case study
Authors: Nilüfer Özdaş; Mehmet Güven Koçak, Serkan Karakış
Abstract: With the widespread use of unmanned aerial vehicles (UAV), high-accuracy photogrammetric mapping studies can be carried out over small areas with cost-effective simple systems. By comparing images obtained at different epochs, 3 Dimensional (3D) change detection studies can be easily performed. Digital surface models (DSM) are obtained from the point cloud (PC) with the processing software, their differences are taken, and temporal changes can thus be modeled. This method is known as DEM (DSM) of Difference (DoD) in practice and has low computational cost. Recently, with the availability and accessibility of powerful computers capable of processing increasing amounts of data, 3D change detection studies can be performed directly with raw PCs without converting them to DSM. Methodologically, DoD and PC-based analysis strategies have different evaluation stages and outputs. With DoD, only changes in the vertical direction can be revealed, while PC comparison methods can produce the 3D change vector. In this study, the well-established DoD method and Multiscale Model-to-Model Cloud Comparison (M3C2), one of the 3D PC comparison methods, were compared. The accuracy of the methods was tested at an active open pit mine site where intensive excavation works have been undertaken. Standard deviation values were found below 11 cm with M3C2 distance and DoD differences obtained from UAV images having average ground sampling distances (GSD) of 5.8-6.9 cm. Only about 1% of the differences were categorized as outliers.
PubDate: Fri, 03 May 2024 00:00:00 +030
- Sentiment analysis from georeferenced social media data using natural
language processing and deep learning: The case of Kahramanmaraş
earthquakes
Authors: Dilan Gözdem Dolu; Alper Şen
Abstract: In natural disaster management, the damages caused by natural disasters can be minimized by using Geographic Information Systems (GIS) in pre-disaster preparation, disaster response and post-disaster recovery stages. The aim of this study is to scrape X (formerly known as Twitter) social media data related to the Kahramanmaraş earthquakes on February 6, 2023 using Selenium and BeautifulSoup libraries in Python programming language, and to examine the post-disaster emotional states of people affected by the earthquake using natural language processing and deep learning methods. Thus, it will be possible to contribute to the planning of the general emotional state of the region at the time of the earthquake and the social and psychological rehabilitation activities to be carried out afterwards in a faster and easier way in the GIS environment. In this study, a sentiment analysis was performed with 87% test accuracy on the scraped and organized dataset from the X platform using the Gated Recurrent Units (GRU) deep network model in natural language processing. In addition, by performing hot spot analysis in the GIS environment, the clustering pattern of the emotional state in X messages related to the earthquake occurred was statistically analyzed. Thus, it was determined that georeferenced social media data of the X platform related to possible earthquakes can be used in sentiment analysis.
PubDate: Fri, 03 May 2024 00:00:00 +030
- Classification of indoor point clouds using machine learning for indoor
mapping
Authors: Sena Varbil; Alper Şen
Abstract: 3-Dimensional point cloud classification of interior spaces is of great importance in the creation of interior models in applications such as indoor mapping, indoor navigation, building renovation, facility management, etc. In this study, point clouds of office rooms in the S3DIS (Stanford 3D Indoor Scene) dataset produced by Stanford University were classified with Random Forest (RF) and Multilayer Perceptron (MLP) machine learning methods, and indoor maps were created. For input data, attributes X, Y, Z and R, G, B were used. The classes include ceiling, floor, wall, door, window, column, table, chair, board, clutter, and bookcase objects. To create indoor maps in the training and test data, the classes were merged as follows: wall, door, window, column, board, and bookcase were merged into one class (merged class-1), and table, chair, and clutter were merged into another class (merged class-2). An office was used for the training data and tested in five different offices. The RF method achieved an average classification accuracy of 88%, and the MLP method achieved an average accuracy of 85%. Thus, indoor maps were obtained with high accuracy, especially thanks to the ceiling and merged class-1, which were classified with high accuracy.
PubDate: Fri, 03 May 2024 00:00:00 +030
- Comparison of ESRI Land Cover and Dynamic World: The case of Cyprus Island
Authors: Ömer Gökberk Narin
Abstract: Monitoring of Land Use/Land Cover (LU/LC) and determination of changes are very important in terms of understanding the relationship between human and the environment. With the development of remote sensing technologies, it has become easier to monitor LU/LC at local and global scales. However, many classification algorithms and methods have been developed and continue to be developed in the classification of remote sensing data. Classification algorithms and methods have advantages and disadvantages against each other. However, the detection of LULC can be used locally and globally. In this study, ESRI Land Cover and Dynamic World data, which are freely available on a global scale, were compared. Both of these data utilise Sentinel-2 imagery for classification and provide 10 m resolution LU/LC data. Cyprus island, an important island in the Mediterranean Sea, is considered in the comparison. For the comparison, firstly the consistency between the two data was analysed. Then, error matrices were created with the control points and their overall accuracy was analysed. There is 95% similarity in the water class, 78% in the crops class, 79% in the built area class, 97% in the trees class, 85% in the bare ground class, and 50% in the flooded vegetation class. Considering the general accuracy, ESRI Land Cover data gave an accuracy of 83.5% while Dynamic World data gave an accuracy of 84.5%. When the results are analysed, it is seen that both data can be used in the monitoring of LULC of the Cyprus island.
PubDate: Fri, 03 May 2024 00:00:00 +030
|
Open Access journal
