Online clustering-based multi-camera vehicle tracking in scenarios with overlapping FOVs

Abstract

Multi-Target Multi-Camera (MTMC) vehicle tracking is an essential task of visual traffic monitoring, one of the main research fields of Intelligent Transportation Systems. Several offline approaches have been proposed to address this task; however, they are not compatible with real-world applications due to their high latency and post-processing requirements. This lack of suitable approaches motivates our proposal: A new low-latency online approach for MTMC tracking in scenarios with partially overlapping fields of view (FOVs), such as road intersections. Firstly, the proposed approach detects vehicles at each camera. Then, the detections are merged between cameras by applying cross-camera clustering based on appearance and location. Lastly, the clusters containing different detections of the same vehicle are temporally associated to compute the tracks on a frame-by-frame basis. The experiments show promising low-latency results while addressing real-world challenges such as the a priori unknown and time-varying number of targets and the continuous state estimation of them without performing any post-processing of the trajectories. Our code is available at http://www-vpu.eps.uam.es/publications/Online-MTMC-Tracking.

Multi-camera Vehicle Tracking and Recognition with Multimodal Contrastive Domain Sharing GAN and Topological Embeddings

“…Vehicle re-identification is particularly useful when tracking vehicles across multiple cameras in a network or when identifying the same vehicle in different video frames or at different times. For vehicle re-identification here TIE-CNN [31] is used for accurate identification of vehicles in multi camera systems. TIE-CNN is a neural network architecture that incorporates topological information into the traditional Convolution Neural Network (CNN) framework.…”

I-24 MOTION: An instrument for freeway traffic science

“…intersection-over-union metric is used to compute affinity between object positions and new detections [92]. IOU is computed based on vehicle footprints in space rather than bounding box coordinates within an image, which allows detections from multiple cameras with distinct fields of view to be incorporated provided accurate homography information is available for each camera (for more information on camera homographies and data coordinate system, see Section IV-B and Appendix C. The multi-camera tracking problem is solved by detection fusion (as in [93], [94]) rather than trajectory fusion (as in [95]) to reduce redundant tracking of the same object in multiple fields of view. Figure 6 shows the result of object detection and tracking within image coordinates, and the corresponding roadway coordinate object positions obtained using image homography.…”