An expert fuzzy system to detect dangerous circumstances due to children in the traffic areas from the video content analysis

摘要

Currently, an important demand for security exists in many fields. In the area of traffic safety, children are vulnerable elements because they may create a dangerous situation easily. Therefore, in this paper, we propose an intelligent surveillance system to detect, in real time, the danger due to the existence of unprotected children in traffic zones. We analyze the behavior of objects from video content analysis. The developed system is based on fuzzy rules for describing semantically the studied risk. The fuzzy logic provides a gradual danger detection. Besides, the model can be adjusted through membership functions to fuzzy concepts. The system is characterized by being highly scalable and flexible. Moreover, it is portable to any environment. We also highlight in our proposal the high-level tracking module developed, which is based on the classification of objects and 3D positioning. This tracking method is robust to errors that 2D tracking can present. The results obtained in the experimental stage show a high system performance.