@article{MTMT:34742506,
	title = {An Automatic Road Surface Segmentation in Non-Urban Environments: A 3D Point Cloud Approach with Grid Structure and Shallow Neural Networks},
	url = {https://m2.mtmt.hu/api/publication/34742506},
	author = {Dowajy, Mohammad and Somogyi, József Árpád and Barsi, Árpád and Lovas, Tamás},
	doi = {10.1109/ACCESS.2024.3372431},
	journal-iso = {IEEE ACCESS},
	journal = {IEEE ACCESS},
	volume = {12},
	unique-id = {34742506},
	issn = {2169-3536},
	abstract = {Automatic road segmentation from three-dimensional point cloud data has gained increasing interest recently. However, it is still challenging to do this task automatically due to the wide variations of roads and complex environments, especially in non-urban areas. This research proposed a comprehensive approach for using shallow neural networks to segment non-urban road point clouds to support autonomous driving applications. The proposed approach involves converting raw point cloud data into a regular grid of cells or partial clouds. Initially, a shallow neural network based on cells’ properties (cell plane fitting error, cell average intensity, cell elevation range, and cell weighted density) was employed to extract road cells from raw point cloud data. The road cells were refined and segmented into inside-road and road border point clouds based on morphologic operations. A point-wise shallow neural network was used to extract road points from the border point clouds based on intensity and geometric features (roughness, curvature, and change rate of the normal). A precise road surface point cloud is obtained by merging the inside-road and filtered border point clouds. The method performance was evaluated for two datasets captured using a mobile laser scanner (MLS). In the first dataset, the road points were extracted at average completeness, correctness, quality, and overall accuracy of 98.40%, 99.13%, 97.56%, and 98.47%, respectively. Similarly, the method achieved high scores for the second dataset with 97.22% completeness, 99.02% correctness, 96.29% quality, and 98.71% overall accuracy. The method performance demonstrates an advancement when compared to various state-of-the-art methods and also confirms its adaptability to different road environments. Authors},
	keywords = {CELLS; NEURAL NETWORKS; NEURAL NETWORKS; Feature extraction; Feature extraction; Cytology; data mining; data mining; NEURAL-NETWORKS; Surface topography; Surface topography; Roads and streets; ROAD; Geometric features; Roads; Three dimensional displays; Three-dimensional displays; features extraction; 3D point cloud; 3D point cloud; Geometric feature; Weighted density; Weighted density; road segmentation; road segmentation; Point-clouds; Three-dimensional display; Point cloud compression; Point cloud compression; shallow neural network; shallow neural network; cell point cloud; plan fitting; Cell point cloud; Plan fitting},
	year = {2024},
	eissn = {2169-3536},
	pages = {33035-33044},
	orcid-numbers = {Somogyi, József Árpád/0000-0002-7247-4470; Barsi, Árpád/0000-0002-0298-7502; Lovas, Tamás/0000-0001-6588-6437}
}

@article{MTMT:34554380,
	title = {A Shallow Neural Network Model for Urban Land Cover Classification Using VHR Satellite Image Features},
	url = {https://m2.mtmt.hu/api/publication/34554380},
	author = {Fawzy, Mohamed and Szabó, György and Barsi, Árpád},
	doi = {10.5194/isprs-annals-X-1-W1-2023-57-2023},
	journal-iso = {ANN PHOTOGRAMMETRY REM SENS SPAT INF SCI},
	journal = {ISPRS ANNALS OF THE PHOTOGRAMMETRY REMOTE SENSING AND SPATIAL INFORMATION SCIENCES},
	volume = {10},
	unique-id = {34554380},
	issn = {2194-9042},
	abstract = {Recently, image classification techniques using neural networks have received considerable attention in sustainable urban development, since their applications have an extreme effect on building distribution, infrastructural networks, and water resource management. In this research, a back-propagation shallow neural network model is presented for very high resolution satellite image classification in urban environments. Workflow procedures consider selecting and collecting data, preparing required study areas, extracting distinctive features, and applying the classification process. Visual interpretation is performed to identify observed land cover classes and detect distinctive features in the urban environment. Pre-processing techniques are implemented to present the used images in a more suited form for the classification techniques. A shallow neural network model (supported by MathWorks MATLAB environment) is successfully applied and results are evaluated. The proposed model is tested for classifying both WorldView-2 and WorldView-3 multispectral images with different spatial and spectral characteristics to check the model's applicability to various kinds of satellite imagery and different study areas. Model outcomes are compared to two well-known classification methods; the Nearest Neighbour object-based method and the Maximum Likelihood pixel-based classifier, to validate and check the model stability. The overall accuracy achieved by the proposed model is 86.25% and 83.25%, while the nearest neighbour approach has obtained 84.50% and 82.75%, and the maximum likelihood classifier has accomplished 82.50% and 80.25% for study area 1 and study area 2 respectively. Obtained results indicate that the developed shallow neural network model achieves a promising accuracy for urban land cover classification in comparison with the standard techniques. © Author(s) 2023. CC BY 4.0 License.},
	keywords = {NEURAL-NETWORKS; BACKPROPAGATION; satellite imagery; Maximum likelihood; water management; Land use; Land use; land cover; land cover; Study areas; Urban environments; Urban environments; Neural network model; Image classification; Image classification; Satellite images; NEURAL NETWORK MODELS; Urban growth; urban land cover classification; Images classification; Shallow neural networks; shallow neural network; VHR satellite images; VHR satellite image},
	year = {2023},
	eissn = {2194-9050},
	pages = {57-64},
	orcid-numbers = {Szabó, György/0000-0001-5280-6143; Barsi, Árpád/0000-0002-0298-7502}
}

@article{MTMT:34554379,
	title = {LINE AND POLYGON TOPOLOGY IN OPENDRIVE MODELLING},
	url = {https://m2.mtmt.hu/api/publication/34554379},
	author = {Lógó, János Máté and Barsi, Árpád},
	doi = {10.5194/isprs-archives-XLVIII-1-W2-2023-835-2023},
	journal-iso = {ISPRS (2002-)},
	journal = {INTERNATIONAL ARCHIVES OF PHOTOGRAMMETRY AND REMOTE SENSING (2002-)},
	volume = {48},
	unique-id = {34554379},
	issn = {1682-1750},
	abstract = {The paper discusses the importance of efficient methods in the automotive industry for the development of self-driving vehicles and advanced vehicle assistants, focusing on the use of high-definition (HD) maps. The integration of computer simulation and HD maps in the OpenDRIVE format is emphasized. A paradigm shift in map topology is highlighted, requiring a new map creation and usage approach. The article presents a methodology that addresses both geometric and topological aspects in creating accurate HD map models. The method focuses on the connection of linear and arc road elements and eliminating continuity/connectivity errors in lane descriptions. Real-world tests validate the implemented methodology and demonstrate the successful generation of topologically correct HD map models. The results show the potential of these models for various automotive applications, particularly in the development and testing of self-driving vehicles and advanced vehicle assistants. The methodology contributes to the advancement of HD map creation, providing valuable insights for researchers and map makers in the automotive industry. © Author(s) 2023.},
	keywords = {Topology analysis; Highly automated driving; HD-MAP; geometric-topologic codesign},
	year = {2023},
	eissn = {2194-9034},
	pages = {835-840},
	orcid-numbers = {Lógó, János Máté/0000-0002-0946-5328; Barsi, Árpád/0000-0002-0298-7502}
}

@article{MTMT:34554377,
	title = {COST-EFFICIENT METHODS OF DERIVING SLOPE INFORMATION FOR ROAD SEGMENTS IN DRIVER-ASSISTANCE APPLICATIONS},
	url = {https://m2.mtmt.hu/api/publication/34554377},
	author = {Horváth, Viktor Győző and Barsi, Árpád},
	doi = {10.5194/isprs-archives-XLVIII-1-W2-2023-895-2023},
	journal-iso = {ISPRS (2002-)},
	journal = {INTERNATIONAL ARCHIVES OF PHOTOGRAMMETRY AND REMOTE SENSING (2002-)},
	volume = {48},
	unique-id = {34554377},
	issn = {1682-1750},
	abstract = {An advanced driver-assistance system (ADAS) is any of a group of technologies that assist drivers in driving and parking functions. Through a safe human-machine interface, ADAS increase car and road safety. These Advanced driver-assistance systems rely on special maps with extended geometry and attribute information. This extra information includes slope, curvature, and speed limit. ADAS-enabled maps are usually rather expensive in the industry. This paper is focused on finding cost-efficient alternatives for generating the slope aspect of ADAS maps. Slope and height information is not only used in ADAS but is a critical aspect of calculating electronic vehicle (EV) ranges, and truck fuel-efficiency calculations as well. ADAS slope information usually requires high-accuracy surveys. This paper researches the possible generation of slope information for road segments with the use of digital elevation models (DEM) or crowdsourcing with low-cost sensors and Kalman filtering. The first approach is based on globally available DEMs with interpolation and filtering with road geometry. DEMs have variable accuracy depending on the type of technology used in producing them. Such technologies include photogrammetry, aerial and terrestrial laser scanning (ALS, TLS), or aerial or space radar measurements. The other method is by using low-cost GPS and IMU sensors for generating altitude profiles. These produced altitude profiles are compared with a profile generated from a high-accuracy survey using large-resolution DEMs produced by aerial photogrammetry or aerial laser scanning. This paper proposes metrics with which these datasets can be compared, one is using the height differences, and the other compares the slope values at discrete common points. In the conclusion, the paper tries to find use cases for the low-accuracy data. © Author(s) 2023.},
	keywords = {GIS; ADAS; crowdsource; slope maps},
	year = {2023},
	eissn = {2194-9034},
	pages = {895-900},
	orcid-numbers = {Horváth, Viktor Győző/0000-0002-4058-2768; Barsi, Árpád/0000-0002-0298-7502}
}

@article{MTMT:34342469,
	title = {Járműfedélzeti kamerák helymeghatározása Kálmán-szűréssel},
	url = {https://m2.mtmt.hu/api/publication/34342469},
	author = {Horváth, Viktor Győző and Barsi, Árpád},
	doi = {10.30921/GK.75.2023.4.2},
	journal-iso = {GEODÉZIA ÉS KARTOGRÁFIA},
	journal = {GEODÉZIA ÉS KARTOGRÁFIA},
	volume = {75},
	unique-id = {34342469},
	issn = {0016-7118},
	year = {2023},
	pages = {15-20},
	orcid-numbers = {Horváth, Viktor Győző/0000-0002-4058-2768; Barsi, Árpád/0000-0002-0298-7502}
}

@CONFERENCE{MTMT:34193828,
	title = {Maps Towards Autonomous Driving Ecosystem},
	url = {https://m2.mtmt.hu/api/publication/34193828},
	author = {Barsi, Árpád and Lógó, János Máté},
	booktitle = {Proceedings of the Seventeenth International Conference on Civil, Structural and Environmental Engineering Computing},
	doi = {10.4203/ccc.6.15.1},
	unique-id = {34193828},
	year = {2023},
	orcid-numbers = {Barsi, Árpád/0000-0002-0298-7502; Lógó, János Máté/0000-0002-0946-5328}
}

@inproceedings{MTMT:34001921,
	title = {Building Extraction from VHR Satellite Images Using Shallow Neural Network model and Object-Based Post- Classification Refinement},
	url = {https://m2.mtmt.hu/api/publication/34001921},
	author = {Mohamed, Fawzy and Szabó, György and Barsi, Árpád},
	booktitle = {Az elmélet és gyakorlat találkozása a térinformatikában XIV. : Theory meets practice in GIS},
	unique-id = {34001921},
	year = {2023},
	pages = {83-90},
	orcid-numbers = {Szabó, György/0000-0001-5280-6143; Barsi, Árpád/0000-0002-0298-7502}
}

@article{MTMT:33698800,
	title = {Evaluation of Topology Description Models in Road Network Formats},
	url = {https://m2.mtmt.hu/api/publication/33698800},
	author = {Lógó, János Máté and Barsi, Árpád},
	doi = {10.12700/APH.20.1.2023.20.10},
	journal-iso = {ACTA POLYTECH HUNG},
	journal = {ACTA POLYTECHNICA HUNGARICA},
	volume = {20},
	unique-id = {33698800},
	issn = {1785-8860},
	abstract = {Topology is a particular feature of road networks. Topology refers to properties, such as the connection of roads, but not only through their axes or reference lines, but also at the level of lanes. The map topology of road networks does not currently have procedures that can be expressed in mathematical formulas (or only to a minimal extent), but can only be implemented by algorithms. Our research, therefore, aimed at observing such topological regularities and then developing a method of investigation, by constructing rules and additional algorithms. To test this work, we used synthetic and real data, focusing on the case of map content embodied in four formats. In this paper, we present the test methods for the data models, the results of our test runs and finally, we point out that topological checks are extensively justified to determine the quality of the produced map. In the future we plan to develop an automatic correction mechanism based on this.},
	keywords = {Road network; Topology analysis},
	year = {2023},
	eissn = {1785-8860},
	pages = {143-156},
	orcid-numbers = {Lógó, János Máté/0000-0002-0946-5328; Barsi, Árpád/0000-0002-0298-7502}
}

@article{MTMT:33609760,
	title = {Building Maps Using Monocular Image-feeds from Windshield-mounted Cameras in a Simulator Environment},
	url = {https://m2.mtmt.hu/api/publication/33609760},
	author = {Szántó, Mátyás and Kobál, Sándor and Vajta, László and Horváth, Viktor Győző and Lógó, János Máté and Barsi, Árpád},
	doi = {10.3311/PPci.21500},
	journal-iso = {PERIOD POLYTECH CIV ENG},
	journal = {PERIODICA POLYTECHNICA-CIVIL ENGINEERING},
	volume = {67},
	unique-id = {33609760},
	issn = {0553-6626},
	abstract = {3-dimensional, accurate, and up-to-date maps are essential for vehicles with autonomous capabilities, whose functionality is made possible by machine learning-based algorithms. Since these solutions require a tremendous amount of data for parameter optimization, simulation-to-reality (Sim2Real) methods have been proven immensely useful for training data generation. For creating realistic models to be used for synthetic data generation, crowdsourcing techniques present a resource-efficient alternative. In this paper, we show that using the Carla simulation environment, a crowdsourcing model can be created that mimics a multi-agent data gathering and processing pipeline. We developed a solution that yields dense point clouds based on monocular images and location information gathered by individual data acquisition vehicles. Our method provides scene reconstructions using the robust Structure-from-Motion (SfM) solution of Colmap. Moreover, we introduce a solution for synthesizing dense ground truth point clouds originating from the Carla simulator using a simulated data acquisition pipeline. We compare the results of the Colmap reconstruction with the reference point cloud after aligning them using the iterative closest point algorithm. Our results show that a precise point cloud reconstruction was feasible with this crowdsourcing-based approach, with 54\% of the reconstructed points having an error under 0.05 m, and a weighted root mean square error of 0.0449 m for the entire point cloud.},
	keywords = {sensors; crowdsourcing; Environmental reconstruction; SfM; Structure-from-motion; automotive simulator},
	year = {2023},
	eissn = {1587-3773},
	pages = {457-472},
	orcid-numbers = {Szántó, Mátyás/0000-0003-1793-147X; Vajta, László/0000-0001-7164-6050; Horváth, Viktor Győző/0000-0002-4058-2768; Lógó, János Máté/0000-0002-0946-5328; Barsi, Árpád/0000-0002-0298-7502}
}

@article{MTMT:33128086,
	title = {TOPOLOGICAL ANOMALY DETECTION IN AUTOMOTIVE SIMULATOR MAPS},
	url = {https://m2.mtmt.hu/api/publication/33128086},
	author = {Barsi, M. and Barsi, Árpád},
	doi = {10.5194/isprs-archives-XLIII-B4-2022-343-2022},
	journal-iso = {ISPRS (2002-)},
	journal = {INTERNATIONAL ARCHIVES OF PHOTOGRAMMETRY AND REMOTE SENSING (2002-)},
	volume = {43-B4},
	unique-id = {33128086},
	issn = {1682-1750},
	abstract = {Autonomous driving went through numerous significant improvements over the past couple of years, including driver assistants that are already capable of executing an increasing number of complex tasks without the need for any human intervention. As a result of these changes, manufacturers are relying more and more on fast, cheap, and often better-quality simulations over real-world tests. To create these environments, the real world needs to be transformed to a digital, high-definition model. HD maps - for example, the XML-based, hierarchic OpenDRIVE format - aim to serve this purpose. The most important element of any realistic map format is the ability to check connectivity on the map in a convenient way, hence the need for topology. In HD maps, the description of junctions poses a significant challenge to the designers of the format, since they are essential yet complex topological elements. The representation of these junctions is still in progress, however, according to our analysis, the use of the current tools in OpenDRIVE can result in anomalies in the map. In the most recent release of OpenDRIVE (version 1.7), road-road and lane-lane connections are described using links consisting of a predecessor and a successor. These however, has to be described multiple times when the junction tag is used, resulting in duplicates in the model which can be easily exploited. Our proposed solution for this issue is the elimination of the junction tag, which not only gets rid of the anomalies without any loss of information, but it also significantly reduces the size of the model. In this paper, a detailed explanation is provided of this issue and the proposed solution with examples using OpenDRIVE models.},
	keywords = {environmental model; remote sensing; autonomous driving; Geography, Physical; HD map},
	year = {2022},
	eissn = {2194-9034},
	pages = {343-348},
	orcid-numbers = {Barsi, Árpád/0000-0002-0298-7502}
}