ACAMAI
This project will provide planners, decision makers, city workers and information consumers with constantly updated, normalized and reliable information about accessibility features, improving the quality of life and the cost rational planning of intervention works in urban areas. We will use GSV labeled datasets, e.g., from the Project Sidewalk and transfer learning, and produce also labeled datasets of downtown Porto. Applying state of the art computer vision and deep learning techniques, we will analyze the performance of the algorithms using the two datasets. We will also use and test the Project Sidewalk, which supports this project, labeling tool, which is being used in several cities in the U.S., and by the city of Amsterdam in Europe. At the same time, the project will adhere to open standards for labeling accessibility, which can then be imported into open frameworks, such as the OpenStreetMap (following, for example, the OpenSidewalks data schema, https://www.opensidewalks.com/). The outcomes of the project will be an open labeled accessibility dataset of Porto, and a platform with a model to analyze and identify accessibility information from GSV imagery. After training and calibrating the model, the platform can be used in other geographical areas. The Braga city council, which also supports this project, is prepared to be part of a future pilot project.
The project runs from 01-03-2023 to 30-09-2024 with a budget of 49.814,20 euros.
DOI: https://doi.org/10.54499/2022.09218.PTDC
The following documents and deliverables are available:
A presentation of the project is given here.
Preliminary tests report.
Menção honrosa na categoria Mobilidade e Logística do 3º Prémio Cidades e Territórios do Futuro, iniciativa da Associação Portuguesa para o Desenvolvimento das Comunicações (APDC), 14 de maio de 2024 (https://premiocidades-apdc.pt/vencedores-2024/)
Presentations:
Presentation at Semana da FCT, University Fernando Pessoa, 21 February 2024.
Publications:
Rui Moreira, Sérgio Moita, Feliz Gouveia, José Manuel Torres, Maria João Guerreiro, Sara Sucena, Diogo Ferreira, Maria Alzira Dinis, “An Object-Based Detection Approach for Automating City Accessibility Constraints Mapping”, 2024 IEEE International Conference on Smart Applications, Communications and Networking, SmartNets 2024, 28 - 30 May 2024, Harrisonburg, Washington DC – USA.
Members:
Alzira Dinis
Ana Paula Rocha (LIACC/UP)
Maria João Guerreiro
Rui Moreira
Sara Sucena
Luís Pinto de Faria
Diogo Ferreira (MSc student)
Sérgio Moita (MSc student)
MSc student 3
Collaborators:
Vítor Silva
Feliz Gouveia
João Santos (BSc student)
Carlos Machado de Sousa (BSc student)
João Teixeira (BSc student)
Documents (Internal use)