Thesis Aplicación AIMES: autonomía para usuarios con discapacidad visual en Metro de Santiago
Loading...
Date
2026-10-21
Authors
Journal Title
Journal ISSN
Volume Title
Program
Ingeniería Civil Telemática
Departament
Campus
Campus Santiago San Joaquín
Abstract
Las personas con discapacidad visual enfrentan diversas dificultades para desplazarse de manera autónoma y segura en entornos urbanos complejos, donde la identificación oportuna de infraestructura y señaléticas resulta fundamental para una navegación sin inconvenientes. En el contexto del Metro de Santiago, elementos como ascensores, pisos podotáctiles, escaleras mecánicas y señaléticas direccionales constituyen referencias esenciales para la orientación y movilidad de estos usuarios. Sin embargo, su identificación depende de la asistencia de terceros o del uso de ayudas convencionales, las cuales presentan limitaciones para entregar información contextual en tiempo real. Con el propósito de contribuir a esta problemática, el proyecto AIMES propone un sistema inteligente de asistencia para personas con discapacidad visual basado en técnicas de visión por computador e inteligencia artificial. La presente memoria aborda el desarrollo del componente de visión computacional encargado de detectar en tiempo real elementos relevantes del entorno, específicamente ascensores, pisos podotáctiles, escaleras mecánicas y señaléticas del Metro de Santiago correspondientes a indicaciones de dirección (más específicamente, a la dirección del andén y salida de la estación). Para ello, se diseñó y construyó un conjunto de datos especializado mediante la recopilación, selección y etiquetado de imágenes representativas del entorno del Metro de Santiago. Posteriormente, se entrenaron y evaluaron modelos de detección de objetos basados en YOLOv11, comparando su desempeño mediante métricas de precisión y rendimiento computacional. Finalmente, el modelo seleccionado fue optimizado para (...).
People with visual impairments face various challenges when moving around independently and safely in complex urban environments, where the timely identification of infrastructure and signage is essential for smooth navigation. In the context of the Santiago Metro, elements such as elevators, tactile paving, escalators, and directional signage serve as essential reference points for these users’ orientation and mobility. However, identifying these elements currently depends on assistance from others or the use of conventional assistive devices, which have limitations in providing real-time contextual information. To address this issue, the AIMES project proposes an intelligent assistance system for people with visual impairments based on computer vision and artificial intelligence techniques. This thesis addresses the development of the computer vision component responsible for detecting relevant elements of the environment in real time—specifically elevators, tactile paving, escalators, and directional signage in the Santiago Metro indicating the direction to the platform and station exit. To this end, a specialized dataset was designed and constructed by collecting, selecting, and labeling images representative of the Santiago Metro environment. Subsequently, object detection models based on YOLOv11 were trained and evaluated, with their performance compared using accuracy metrics and computational efficiency. Finally, the selected model was optimized for (...).
People with visual impairments face various challenges when moving around independently and safely in complex urban environments, where the timely identification of infrastructure and signage is essential for smooth navigation. In the context of the Santiago Metro, elements such as elevators, tactile paving, escalators, and directional signage serve as essential reference points for these users’ orientation and mobility. However, identifying these elements currently depends on assistance from others or the use of conventional assistive devices, which have limitations in providing real-time contextual information. To address this issue, the AIMES project proposes an intelligent assistance system for people with visual impairments based on computer vision and artificial intelligence techniques. This thesis addresses the development of the computer vision component responsible for detecting relevant elements of the environment in real time—specifically elevators, tactile paving, escalators, and directional signage in the Santiago Metro indicating the direction to the platform and station exit. To this end, a specialized dataset was designed and constructed by collecting, selecting, and labeling images representative of the Santiago Metro environment. Subsequently, object detection models based on YOLOv11 were trained and evaluated, with their performance compared using accuracy metrics and computational efficiency. Finally, the selected model was optimized for (...).
Description
Keywords
Discapacidad visual, Accesibilidad, Visión por computador, Inteligencia artificial, Detección de objetos, Aplicación móvil
