Adaptive tracking method for sea-surface target scale of unmanned ship in high sea state based coherent filtering

Abstract

The invention provides an adaptive tracking method for the sea-surface target scale of an unmanned ship in the high sea state based coherent filtering, especially based on a kernel coherent filteringKCF algorithm, and aims at solving the problem of high seat state level, wide target scale change and low resolution in common sea-surface target tracking of the water surface unmanned ship. In the high sea state, sea skyline detection helps realize adaptive displacement and zooming of a target search area, alleviate a boundary effect of coherent filtering, and cope with severe jittering of a target in the scene; and optimization is made in the aspects including target preprocessing, characteristic selection and scale estimation to improve the robustness of the algorithm and solve problems including change of the target scale and low resolution. Compared with coherent filtering algorithms that can meet the real-time tracking requirement of the unmanned ship in recent years, it is shown that the algorithm of the invention has advantages in the aspect of sea-surface target tracking, especially in the high sea state.

Description

technical field

[0001] The invention relates to a correlation filter-based self-adaptive tracking method for a sea surface target scale of an unmanned ship under high sea conditions, belonging to the field of computer vision target tracking. Background technique

[0002] The surface unmanned vehicle is an intelligent surface carrier platform integrating environmental perception, mission planning, motion planning, behavior execution and other modules. It can realize path planning, autonomous navigation and obstacle avoidance, target detection, identification and tracking, etc. A series of civil and military tasks such as marine water quality testing, underwater archaeology, ocean cruising, and ship escort. Among them, the environment perception module realizes the acquisition of three-dimensional environmental information around the unmanned boat by carrying sensors such as cameras and lidars, which is a prerequisite for completing the functions of other modules. The visual ...

Images