Crowd Behavioral Analysis in Video Surveillance

-With the rapid advances of wireless communication and positioning technologies in mobile systems, the acquisition of spatio-temporal data using mobile devices is becoming pervasive. Many applications, such as traffic control systems, geographical information systems, and location-aware advertisement can benefit from efficient processing of spatio-temporal queries. Early methods proposed to efficiently process spatio-temporal queries focus exclusively on Euclidean spaces (i.e., the query results are determined based on the Euclidean distance between each moving object and the object issuing the spatio-temporal query). However, in most real-world applications, the movements of objects (e.g., cabs and pedestrians) are constrained to a transportation network. As a result, the distance between two objects should be computed based on the connectivity of the network rather than the two objects’ locations so that the query results obtained from performing the early methods are not always useful. Recently, several studies have investigated how to process the spatial-temporal queries in road networks, where the criterion for determining the query results is the shortest network distance (i.e., shortest path) between objects. However, the focus of these studies is on providing efficient algorithms to process the K-Nearest Neighbor (KNN) and range queries over moving objects.