FII's Master students WADe Traffic Sign Smart Detector project
Having several (snapshots of) video recordings – captured via a Webcam or uploaded by a user – regarding an urban route (frequently/randomly) followed by a person or a group of persons (e.g., by using a bike/car/bus), develop a (micro-)service-based Web system able to detect road/traffic signs marking this route (road, highway). This detection process could be performed automatically by using specific public APIs and/or by using user-reported info (for example, by using external crowdsourcing navigation services like Waze or alternatives). An ontology specified in OWL will be created and/or adapted to specify things of interest (mainly, a classification of road/traffic signs and their meanings and legal interpretations). For each recognized (category of) road sign, a SPARQL endpoint will offer various knowledge: meaning, type, legal regulations, relationships to other traffic signs, practical advices, context of use, comparisons, plus suggestions regarding user (driver/pedestrian) behavior.
Study the Comparison of European road signs.
Additional resources:
- Traffic Sign Detection Articles @ Google Scholar
- Traffic Sign Recognition Code Repositories @ GitHub
- Denis Aenasoaei - [email protected]
- Maria Istrate - [email protected]
- Cristi Rusu - [email protected]
For solving the presented task we are going to build a Web Application based on microservices. The main components of the application will be:
- C#/.NET component that will offer endpoints for connecting the other components of the application and the database.
- Video Recognition and Image Processing using Python, OpenCV and Tensorflow.
- Angular 10 Single Page App containing a Data Presenter, Upload Service and API Caller.
- Various services for crawling application like Google Maps, Waze, Road Sign DB with the purpose of gathering informations regarding user routes or traffic signs.
- A database for storing traffic sign informations compiled from the users.