A Multisensory Monitoring and Interpretation Framework Based on the Model–View–Controller Paradigm - Foundations on Natural and Artificial Computation - page 448

Activity Detection: The higher level of the framework is devoted to the detec-
tion of activities from the events detected in the previous levels. As the events are
generated by remote (distributed) nodes, their fusion provides a general view of
the activities carried out in the scenario. This level provides the common model
with high level activities detected that may involve several sensors in the scenario
and several objects.
Scenario Modeling: Even though scenario modeling does not appear as a
level within the framework definition, it is a key aspect that enables the rela-
tions among sensor information for situating the objects in the scenario. Scenario
modeling can be seen as a two stage modeling. On the one hand, there is a global
modeling that includes aspects such as map definition, map calibration, environ-
mental condition setting (ambient light or temperature), and sensor placement
and its range. On the other hand, there exists a modeling that is sensor-specific
like the camera field of view. Some parameters are dynamically updated accord-
ing to the sensed values. Table 2 summarizes the main features of the architecture
levels, as well as the parameters provided to the common model.
5 Conclusions
This paper has introduced a monitoring and interpretation framework inspired
in the MVC paradigm. The paper has proposed the extension of the traditional
MVC paradigm for the inclusion of the functionalities of a monitoring and inter-
pretation system. The proposed model is defined as a hybrid distributed system
where remote nodes perform lower level processing as well as data acquisition.
A a central node is in charge of collecting the information and of its fusion.
The remote nodes hold the following layers of the framework: acquisition, sensor
fusion, localization and filtering, localization and filtering fusion, blob detection,
object identification, object classification, object tracking, event detection, event
fusion. The central nodes incorporates activity detection and scenario modeling.
Acknowledgements
This work was partially supported by Spanish Ministerio de Ciencia e Inno-
vacion TIN2010-20845-C03-01 and TIN2010-20845-C03-02 grants, and by Junta
de Comunidades de Castilla-La Mancha PII2I09-0069-0994 and PEII09-0054-
9581 grants.
References
1. Celik, T., Kusetogullari, H.: Solar-powered automated road surveillance system
for speed violation detection. IEEE Transactions on Industrial Electronics 57(9),
3216-3227 (2010)
2. Fernandez-Caballero, A., Castillo, J.C., Serrano-Cuerda, J., Maldonado-Bascon,
S.: Real-time human segmentation in infrared videos. Expert Systems with Appli-
cations 38(3), 2577-2584 (2011)