Multi-feature fusion-based robust tracking of small targets in unmanned surface vehicle vision

Abstract

Objectives To overcome the challenges of tracking small targets in unmanned surface vehicle vision under the conditions of low feature resolution and similar environmental information, a multi-feature fusion-based continuous convolution operator tracking （MCCOT） algorithm is proposed. Methods The resolution of multi-feature maps is enhanced using bicubic interpolation techniques to enable sub-pixel-level localization. Efficiencies in target tracking are achieved through feature projection and sample space generation to mitigate filter overfitting. Furthermore, interference arising from similar environmental features on the filter is addressed by developing an update strategy for high-confidence models. Results As the experimental results show, compared to traditional continuous convolution operator tracking algorithms, the proposed algorithm achieves an average success rate increase of 17.4%, average distance precision increase of 17.8%, and expected average overlap rate increase of 5.1%. Conclusions The proposed algorithm can deal with the problem of small target tracking confusion in marine environments, providing key technical support for improving the intelligent sensing capability of unmanned boats and marine robots.

Keywords

• robust tracking of small targets at sea
• multi-feature fusion
• continuous convolutional operator
• unmanned surface vehicle vision