dc.contributor | Bulla Cruz, Lenin Alexander | |
dc.contributor | Mangones Matos, Sonia Cecilia | |
dc.contributor | Grupo de Investigación en Logística para el Transporte Sostenible y la Seguridad - TRANSLOGYT | |
dc.creator | Acosta Sequeda, Juan Guillermo | |
dc.date.accessioned | 2021-08-13T14:59:19Z | |
dc.date.available | 2021-08-13T14:59:19Z | |
dc.date.created | 2021-08-13T14:59:19Z | |
dc.date.issued | 2021-07 | |
dc.identifier | https://repositorio.unal.edu.co/handle/unal/79946 | |
dc.identifier | Universidad Nacional de Colombia | |
dc.identifier | Repositorio Institucional Universidad Nacional de Colombia | |
dc.identifier | https://repositorio.unal.edu.co/ | |
dc.description.abstract | Elementary units represent an accurate approach to quantifying road exposure in a way that it acquires statistical meaning as trials with possible outcomes. This enables the possibility of conducting sophisticated statistical analysis in a field that allows planners and policy makers to make decision without waiting for people to die to have useful data. This potential analysis has more value if the amount of data available is sufficiently big. However, manually extracting this data from on-site or video observations is a very difficult and time consuming task. Automated traffic video analysis tools allow not only the faster gathering of data but also the standardization and re-productivity of the work conducted. This thesis proposes
and automatic video estimator of road exposure by means of vehicle detection based on a convolutional neural network. The resulting algorithm is tested in three different intersections with increasing levels of difficulty in terms of camera angle, traffic volumes, road users, and occlusions. As a result, confusion matrices for each intersection were obtained with their respective F1 scores, which indicated that the intersection thought to be the middle one in level of difficulty ended up showing the best performance of the algorithm. Fisher’s Exact statistical test was also computed in order to test the manual and automatic distribution counts correspondence. The different variables affecting the algorithm such as angles, user input parameters, and the apparent size of vehicles are discussed, and from that point the scope of future research is formulated. (Text taken from source) | |
dc.description.abstract | Las medidas elementales de exposición constituyen una aproximación precisa a la cuantificación de la exposición vial, de tal manera que esta adquiere significado estadístico en la forma de pruebas con distinto resultados posibles. Lo anterior, posibilita llevar a cabo análisis estadísticos sofisticados en un campo que permite a planificadores y trabajadores en políticas públicas el tomar decisiones sin tener que esperar a que las personas mueran para tener datos útiles. Este análisis potencial tiene aun más valor si la cantidad de datos es lo suficientemente grande. Sin embargo, extraer esta información de forma manual en campo o a partir de videos es una tarea difícil y dispendiosa. Las herramientas de análisis automático por video permiten no solo recolectar información de forma más rápida sino también la estandarización y reproducibilidad del trabajo. Esta tesis propone una forma automática de estimar la exposición vial por medio de video a través de la detección de vehículos basada en una red neuronal convolucional. El algoritmo resultante es puesto a prueba en tres intersecciones viales diferentes y con nivel de dificultad incremental en términos de ángulos de grabación, volúmenes de tráfico, usuarios viales y oclusiones. Como resultado, se obtienen las matrices de confusión de cada intersección con sus respectivos score F1, que indicaron que la intersección que se consideraba de nivel moderado de dificultad fue en realidad la que presentó el mejor desempeño. El test exacto de Fisher fue empleado para determinar la correspondencia entre la distribución de conteos de eventos manuales y automáticos. Las distintas variables que afectan el funcionamiento del algoritmo, tales como ángulos, parámetros de usuario y el tamaño aparente de los vehículos son discutidos en detalle y, a partir de estos, se propone la ruta para futuras investigaciones. (Texto tomado de la fuente) | |
dc.language | eng | |
dc.publisher | Universidad Nacional de Colombia | |
dc.publisher | Bogotá - Ingeniería - Maestría en Ingeniería - Transporte | |
dc.publisher | Facultad de Ingeniería | |
dc.publisher | Bogotá, Colombia | |
dc.publisher | Universidad Nacional de Colombia - Sede Bogotá | |
dc.relation | Alom, M. Z., Taha, T. M., Yakopcic, C., Westberg, S., Sidike, P., Nasrin, M. S., Hasan, M., Van Essen, B. C., Awwal, A. A. & Asari, V. K. (2019), ‘A state-of-the-art survey on deep learning theory and architectures’, Electronics (Switzerland) 8(3). | |
dc.relation | Bulla-Cruz, L. A., Laureshyn, A. & Lyons, L. (2020), ‘Event-based road safety assessment: A novel approach towards risk microsimulation in roundabouts’, Measurement: Journal of the International Measurement Confederation 165, 108192. URL: https://doi.org/10.1016/j.measurement.2020.108192 | |
dc.relation | Chapman, R. (1973), ‘The concept of exposure’, Accident Analysis and Prevention 5(2), 95– 110. | |
dc.relation | Elvik, R. (2015), ‘Some implications of an event-based definition of exposure to the risk of road accident’, Accident Analysis and Prevention 76(0349), 15–24. | |
dc.relation | Elvik, R., Erke, A. & Christensen, P. (2009), ‘Elementary Units of Exposure’, Transportation Research Record: Journal of the Transportation Research Board 2103(1), 25–31. | |
dc.relation | European Comission (2015), ‘InDeV — Innovation and Networks Executive Agency’. URL: https://ec.europa.eu/inea/en/horizon-2020/projects/h2020-transport/safety/indev | |
dc.relation | Forero, A. & Calderon, F. (2019), ‘Vehicle and pedestrian video-tracking with classification based on deep convolutional neural networks’, 2019 22nd Symposium on Image, Signal Processing and Artificial Vision, STSIVA 2019 - Conference Proceedings pp. 1–5. | |
dc.relation | Hauer, E. (1982), ‘Traffic conflicts and exposure’, Accident Analysis and Prevention 14(5), 359–364. | |
dc.relation | Hauer, E. (1995), ‘On exposure and accident rate’, Traffic engineering and control 36(3), 134– 138. | |
dc.relation | Holmberg Bahnsen, C. & Bornø Jensen, M. (2016), Title: RUBA-video analysis software for road user behaviour analyses, Technical report. | |
dc.relation | Hopfield, J. J. (1982), ‘Neural networks and physical systems with emergent collective computational abilities.’, Proceedings of the National Academy of Sciences of the United States of America 79(8), 2554–2558. URL: https://www.pnas.org/content/79/8/2554 https://www.pnas.org/content/79/8/2554.abstract | |
dc.relation | Hurtik, P., Molek, V., Hula, J., Vajgl, M., Vlasanek, P. & Nejezchleba, T. (2020), ‘Poly- YOLO: Higher speed, more precise detection and instance segmentation for YOLOv3’,arXiv (May). | |
dc.relation | Johnsson, C., Laureshyn, A., D ́agostino, C. & De Ceunynck, T. (2020), ‘The ‘safety in density’ effect for cyclists and motor vehicles in Scandinavia: An observational study’, IATSS Research pp. 4–10. URL: https://doi.org/10.1016/j.iatssr.2020.08.003 | |
dc.relation | Johnsson, C., Nor ́en, H. & Laureshyn, A. (2018), ‘T-Analyst - semi-automated tool for traffic conflict analysis’, InDev, Horizon 2020 project (Deliverable 6.1). | |
dc.relation | Kohonen, T. (1982), ‘Self-organized formation of topologically correct feature maps’, Biolo- gical Cybernetics 43(1), 59–69. URL: https://link.springer.com/article/10.1007/BF00337288 | |
dc.relation | Laureshyn, A., Goede, M. d., Saunier, N. & Fyhri, A. (2017), ‘Cross-comparison of three surrogate safety methods to diagnose cyclist safety problems at intersections in Norway’, Accident Analysis and Prevention 105, 11–20. URL: http://dx.doi.org/10.1016/j.aap.2016.04.035 | |
dc.relation | Lin, T. Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., Doll ́ar, P. & Zitnick, C. L. (2014), Microsoft COCO: Common objects in context, Technical Report PART 5. Macukow, B. (2016), ‘Information Density Based Image Binarization’, Springer International Publishing Switzerland 1, 105–115. | |
dc.relation | McCulloch, W. & Pitts, W. (1943), ‘A logical calculus of the ideas immanent in nervous activity’, Bulletin of mathematical biophysics 5, 115–133. | |
dc.relation | Mensah, A. & Hauer, E. (1998), ‘Two Problems of Averaging Arising in the Estimation of the Relationship Between’, Transportation Research Record 1(98), 37–43. | |
dc.relation | Persaud, B. N. & Mucsi, K. (1995), ‘Microscopic accident potential models for two-lane rural roads’, Transportation Research Record (1485), 134–139. | |
dc.relation | Redmon, J., Divvala, S., Girshick, R. & Farhadi, A. (2015), ‘You Only Look Once: Unified, Real-Time Object Detection’, Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition 2016-December, 779–788. URL: http://arxiv.org/abs/1506.02640 | |
dc.relation | Risk, A. & Shaoul, J. E. (1982), ‘Exposure to risk and the risk of exposure’, Accident Analysis and Prevention 14(5), 353–357. | |
dc.relation | Rosebrock, A. (2018), ‘YOLO object detection with OpenCV’. | |
dc.relation | Rumar, K. (1988), ‘Collective risk but individual safety’, Ergonomics 31(4), 507–518. | |
dc.relation | Rumar, K. (1999), Road safety and benchmarking, in ‘Proceedings of the Paris Conference on Transport Benchmarking’, pp. 95–109. | |
dc.relation | Rumelhart, D. E. & McClelland, J. L. (1987), Learning Internal Representations by Error Propagation - MIT Press books, in ‘Parallel Distributed Processing: Explorations in the Microstructure of Cognition: Foundations’, MIT Press, pp. 318–362. URL: https://ieeexplore.ieee.org/document/6302929 | |
dc.relation | Saha, S. (2018), ‘A Comprehensive Guide to Convolutional Neural Networks — the ELI5 way’. | |
dc.relation | Salvador, S. & Chan, P. (2007), ‘Toward accurate dynamic time warping in linear time and space’, Intelligent Data Analysis 11(5), 561–580. | |
dc.relation | Saunier, N. & Midenet, S. (2010), ‘Automatic Estimation of the Exposure to Lateral Collision in Signalized Intersections using Video Sensors’, arXiv preprint arXiv:1012.4776 1(514), 1– 9. | |
dc.relation | Transport Systems (2017), ‘Deodata - Recolecci ́on de informaci ́on de tr ́ansito mediante el uso de videos’. URL: https://www.deodata.co/ | |
dc.relation | WHO (2018), ‘Global Status Report on Road Safety 2018’, WHO . | |
dc.rights | Atribución-NoComercial-SinDerivadas 4.0 Internacional | |
dc.rights | http://creativecommons.org/licenses/by-nc-nd/4.0/ | |
dc.rights | info:eu-repo/semantics/openAccess | |
dc.rights | Derechos reservados al autor, 2021 | |
dc.title | A computational tool for the automatic detection of exposure to traffic risk from elementary events | |
dc.type | Trabajo de grado - Maestría | |