Implementation of safety monitoring system at the waterfront based on video analysis technique

There are many object detection and tracking algorithms and mechanisms that have been developed for various environments and for various reasons, all with their strengths and weaknesses. A real-life application of object detection should be as fast as possible, and as lightweight as possible given the circumstances of many people in a single area increasing the computational cost. The emphasized lightweight method that's been developed in this paper is composed of two stages: First, the pedestrian detection is done in two steps, initially with Haar-like features of Adaboost for full body detection, and Local Binary Patterns (LBP) for head detection. The chosen tracking algorithm is Probabilistic Data Association Filter (PDAF) for its robustness and lightweight nature. The developed system is used for coastline safety near tetrapods against emergency situations.