Executive Summary: This project contributes to the long-term goal of obtaining continuous data on the use of road infrastructure and public space. The municipalities carry out automated counts of vehicles on avenues, but pedestrians and cyclists are not registered. Neither are open data approaches used, so the information collected is not easily and freely accessible by citizens. In addition, manual counts are expensive and years can pass between counts.
The project aims to acquire knowledge in the area of computer vision (CV) through the replication and improvement of a prototype of a managed video camera and all the infrastructure at the server level necessary to publish the data obtained by the software in a standardized manner. computer vision to generate a product that automatically performs the volumetric count of pedestrians, cyclists, cars and buses to characterize the use of public space thanks to space-time measurements.
The main objective is to improve the existing software of the prototype so that it automatically performs volumetric counts of the mobility actors, incurring an acceptable error.
The improvements imply the use and systematic evaluation of the parts and modules used by the integrated controller (embedded) in the video camera prototype, which performs the segmentation, classification, and communication of the results to a remote server; using techniques such as: Gradient Histogram (HoG), Support Vector Machines (SVM) and other supervised learning techniques.
In addition, it seeks to integrate the measurements in a remote platform that allows participation and consultation of the data generated by the community on the number of users according to the time of day. The use of Spatial Data Infrastructure (IDE) and protocols for handling time series is not ruled out, see for example the OGC-SOS protocols. The use of Processing Services on the WEB (OGC-WPS) for the automated generation of reports is also analyzed.
A great finding would be to provide tools to empower citizens on issues of mobility and use of public space. There are relationships with the National Plan for Good Living (PNBV) in goal 6.1 to detect lack of protection infrastructure. Citizen participation and involvement make it possible to meet goal 5.1. Similarly, studying the manufacture of sensors in the country contributes to goals 10.3 and 10.4. The results have implications for tourism, that is, with goal 10.8.
These Results are important due to the use of current technologies in topics of general interest, encouraging citizens to take an interest in learning about and improving the country's situation.
Objectives General: It must be achieved during the development of the project. It identifies the purpose towards which resources and efforts should be directed. It is the set of results that the project intends to achieve through the actions.
Implement an open information system that includes computer vision algorithms and techniques, an adequate information exchange system, and a front-end for information presentation to contribute to the management of public/private space.
Specific objectives:
- Develop and/or improve computer vision techniques to automate volumetric counts of people, cyclists, and cars.
- Develop information sharing strategies to lessen the impact of data collection on the source device.
- Define a method of data presentation through an open and public information system.
Participating Institutions:
UDLA, UTPL, UTMACH.
Participants:
Project Director Dr. Gabriel Barros.
- gabriel barros
- Angel Valdivieso
- Omar Ruiz
- louis beard
- Nancy Loja
- wilmer rivas
Awarded budget: $32400
Project status: Signing of agreements.
