Qtrajectories: Improving the Quality of Object Tracking Using Self-Organizing Camera Networks

Abstract

Previous work in the research field of video surveillance intensively focused on separated aspects of object detection, data association, pattern recognition and system design. In contrast, we propose a holistic approach for object tracking in a self-organizing and distributed smart camera network. Each observation task is represented by a software-agent which improves the tracking performance by collaborative behavior. An object tracking agent detects persons in a video stream and associates them with a trajectory. The pattern recognition agent analyses these trajectories by detecting points of interest within the observation field. These are characterized by a non-deterministic behavior of the moving person. The trajectory points (enriched by the results of the pattern recognition agent) will be used by a configuration agent to align the cameras field of view. We show that this collaboration improves the performance of the observation system by increasing the amount of detected trajectory points by 22%.

SmaCCS: Smart Camera Cloud Services - Towards an Intelligent Cloud-based Surveillance System

“…Smart Camera System. SmaCCS builds upon the results of the CamInSens system (D'Angelo et al, 2012) and QTrajectories project (Jänen et al, 2012). Fig.…”

“…Therefore, video-data from several SCs needs to be aggregated and suspicious movements (in terms of trajectories, cf. the QTrajectories project that serves as input (Jänen et al, 2012)) have to be detected. Additionally, the movements and suspicious events are categorised by a pre-trained alarmclassification system.…”

“…Thereby, the Cloud-solution serves as interface to the user and as basis for computation-intensive tasks like video analysis (e.g. detection of persons and movements, see (Jänen et al, 2012)), pattern recognition, and learning of conspicuous and abnormal behaviour. A possibly large set of SCs will most probably result in a huge amount of data that needs specific analyses methods following the Big Data principle (Bryant et al, 2008), which is most promisingly tackled using a Cloud-based approach.…”

Persistent Object Tracking With Randomized Forests

“…In such applications, where tracking results support autonomous operation of a system, e.g. in Jaenen et al (2012), speed and the robustness of the data association strategy for linking detections to targets, is of crucial importance. Traditionally, data association is based on geometry and appearance based similarity cues.…”