Outdoor Vacant Parking Space Detector for Improving Mobility in Smart Cities

Cognitive surveillance architecture for scenario understanding

“…Examples of simple and complex events (where 'sth' stands for something), as well as situations resulting from the aggregation of complex events .. 25. Comparison of accuracy rates/final system response when feedback strategies are integrated in the operational workflow of a selected baseline system [Bravo13]. (*) In this case, due to the constant illumination conditions, there is no sudden illumination transition so that the repair-oriented feedback strategy is never launched by the adapted system ..…”

“…Exploiting the capabilities that Computer Vision offers, an alternative to those ITS commercial solutions for parking space detection that rely on other sensors different from cameras is presented in [Bravo13]. The system is able to detect vacant spaces and classify them by the type of vehicle that could park in that area.…”

“…The architecture of the baseline system is depicted in the following Figure 41. Reader is invited to consult [Bravo13] for more information about the different modules this system is composed of as well as the state of the art algorithms each module relies on. Ƈ 1 ; ; which collects a priori information about the set of contextual objects and target objects whose behaviour is analysed for scene understanding purposes:…”

“…Once vehicle and parking spaces are duly modelled, the object manager is in charge of launching process in charge of Dissimilarity Graph analysis that, as represented in the following figure, is a suitable algorithm for detecting vehicles within a parking zone. More details can be found in [Bravo13]. At a higher level of abstraction, the structural model that defines the set of information to be delivered by modules in charge of the situation analysis is:…”

“…Thanks to the introduction of a new level of abstraction, L3, performance of the outdoor vacant detector system in [Bravo13] can be reinforced [Sánchez14]. As introduced in Chapter 4, objective behind the application of feedback strategies as part of abstraction level L3, is to let the system decide if the processing stage could provide additional information by executing an additional routine (e.g.…”