A SURVEY ON ANOMALOUS MOTION DETECTION IN VIDEO DATA

Abstract

The amount of data that has to be processed increases along with the demand for labor to understand the data due to substantial reduction in the cost of transmitting and storing video data. An important component of a video surveillance system is anomaly detection. Because manually analyzing and monitoring such a big amount of video would be impossible. For automated video analysis to determine typical and aberrant vehicle motion behavior for traffic management, public services, and law enforcement, automatic tracking is a crucial necessity. Even though a few anomaly detection techniques have been proposed in the past, there are still a limited number of them that can handle illumination variations, smaller object detection and tracking in poses that change, learning the movement of heterogeneous objects, anomalies in sparse and dense conditions, and tracking under occlusion that the human eye can easily detect. This survey study provides an overview of numerous techniques that have been employed to find anomalies in diverse applications.