Cross-correlation-based robust object tracking in aerial videos

Date

2019

Authors

Perera, A.G.
Law, Y.W.
Chahl, J.

Editors

Advisors

Journal Title

Journal ISSN

Volume Title

Type:

Conference paper

Citation

AIAC18: 18th Australian International Aerospace Congress (2019): HUMS - 11th Defence Science and Technology (DST) International Conference on Health and Usage Monitoring (HUMS 2019): ISSFD - 27th International Symposium on Space Flight Dynamics (ISSFD), 2019, pp.222-228

Statement of Responsibility

Conference Name

AIAC18: 18th Australian International Aerospace Congress (2019) (24 Feb 2019 - 28 Feb 2019 : Melbourne: Engineers Australia, Royal Aeronautical Society)

Abstract

Normalized cross-correlation (NCC) is a well-known technique in visual feature tracking. However, it is sensitive to the scale, rotation and warping differences in the target object. In this work, we propose a method to use with the NCC filter for rotation- and scale-invariant object tracking. The proposed solution consists of three modules: (i) multiple appearance generation in the search image at different rotation angles and scales, (ii) bounding box drift correction by a re-initialization step, and (iii) failure handling by tracker combination. A point tracker that uses the Kanade-Lucas-Tomasi feature-tracking algorithm and a histogram-based tracker that uses the continuously adaptive mean shift (CAMShift) algorithm are used as supporting trackers. A "Faster R-CNN" detector is used to adjust the estimated bounding boxes. The proposed combined tracker is evaluated using the VisDrone 2018 and VOT2018 datasets and compared with five recently published trackers.

School/Discipline

Dissertation Note

Provenance

Description

Access Status

Rights

Copyright 2019 Engineers Australia Access Condition Notes: Accepted manuscript is available open access

License

Grant ID

Call number

Persistent link to this record