Calibration and object correspondence in camera networks with widely separated overlapping views. Issue 3 (1st June 2015)

• Record Type: Journal Article
• Authors: Yin, Fei; Makris, Dimitrios; Velastin, Sergio A.; Ellis, Tim
• Is Part Of: IET computer vision. Volume 9:Issue 3(2015)
• Page Start: 354
• Page End: 367
• Publication Date: 2015-06-01
• Subjects: calibration -- object tracking -- feature selection -- pedestrians -- road vehicles -- image segmentation -- video surveillance -- video cameras; object correspondence -- widely separated overlapping views -- multicamera object tracking -- visual surveillance -- semiautomatic scene calibration method -- camera network mapping -- single ground plane view -- semisupervised approach -- user-selected line scene features -- homography recovery -- camera views -- scene calibration information -- intersection points -- projected vertical axis -- single camera blobs -- pedestrians -- vehicles -- segmentation noise -- occlusions -- quantitative tracking performance evaluation; Computer vision -- Periodicals; Pattern recognition systems -- Periodicals; 006.37
• DOI: 10.1049/iet-cvi.2013.0301
• Languages: English
• ISSNs: 1751-9632

Abstract

Abstract : This study contributes in two ways to the research of multi‐camera object tracking in the context of visual surveillance. Firstly, a semi‐automatic scene calibration method is proposed to deal with mapping a network of cameras with overlapped fields of view onto a single ground plane view, even when the overlap is not substantial. The proposed method uses a semi‐supervised approach that combines tracked blobs with user‐selected line scene features to recover the homographies between camera views that are both simple and accurate. Then, based on the scene calibration information, the intersection points of the projected vertical axis of single camera blobs are used to make object correspondences across multiple views. The method works in mixed environments of both pedestrians and vehicles, and is shown to be accurate and robust against segmentation noise and occlusions. Finally, the advantage of the proposed method is demonstrated by quantitative tracking performance evaluation and comparison against previous methods.