June 7, 2024DronesOpen Access

Visual Object Tracking Based on the Motion Prediction and Block Search in UAV Videos

View Full Paper

Ask AI

Bookmark

Authors

Discussion

Proposed a tracking algorithm that improves performance metrics in UAV videos, suggesting effective long-term tracking solutions.

Achieving a 19.1% increase in success rate and a 20.8% boost in accuracy highlights significant advancements in visual object tracking.
Extensive experiments on five aerial benchmarks validate the effectiveness of the proposed motion prediction and block search algorithm.
The method employs a Kalman filter for motion prediction and a block search module to enhance tracking accuracy in UAV footage, reducing drift substantially when obstacles arise. The introduced dynamic template updating network adapts to changes—empowering the tracker under variable conditions.

Sun et al. (2024) studied this question.

View Full Paper

Ask AI

Bookmark