Industrial Supervision System based on Visual Data Mining and Motion Trajectory Analysis
| Authors: | Fuertes Martínez Juan José, Universidad de León, Spain
Reguera Acevedo Perfecto, Universidad de León, Spain
Dominguez González Manuel, Universidad de León, Spain
Díaz Blanco Ignacio, Universidad de Oviedo, Spain
Cuadrado Vega Abel Alberto, Universidad de Oviedo, Spain |
|---|
| Topic: | 6.4 Safeprocess |
|---|
| Session: | Signal Based Fault Detection and Isolation |
|---|
| Keywords: | Supervision, scene analysis, Man-Machine Interfaces, neural networks, Self-Organizing system, monitoring, residues, visual surveillance, complex systems. |
|---|
Abstract
The current trends in video surveillance systems aim to incorporate mechanisms that understand and remember the activity in a scene to make decisions. These decision modules have as input the object trajectories in a scene resulting from the treatment of the images captured by a video camera. In this paper we propose a novel industrial supervision system for complex multivariable processes that incorporates this decision module. The scenario in this case does not come from the treatment of a video image sequences but from the projection of process variables on a 2D plane using dimension reduction techniques.
