PRODUCTION OF DIGITAL ROAD MAPS BY CROWDSOURCING

Computer systems and methods for updating and/or supplementing a digital road map through crowdsourcing, based on the generalization of geolocation systems that are integrated in the majority of modern road vehicles. The signals collected by these geolocation systems are used to update and/or supplement a digital road map through crowdsourcing. The collected data make it possible to extract data from geographical traces associated with vehicles traveling the road network and: extracting, for each geographical trace, a trajectory curve passing substantially through all of the measurements of the geographical trace; detecting the inflection points (vertices) of each trajectory curve; grouping together all of the vertices into a plurality of vertex classes, using an unsupervised classification algorithm; selecting the most central vertex in each vertex class (representative); forming, from each geographical trace, a road segment between representatives that successively intersect the course of the geographical trace when they are considered in pairs.