Research on Robustness Tracking of Maneuvering Target for Bionic Robot

다국어 초록 (Multilingual Abstract)

In order to improve vision tracking quality of the bionic robot, the new automatic tracking algorithm system is proposed in this paper. Base on design of FPGA image acquisition system, the scene noise is removed by adaptive wiener filtering. Aiming at the problem of ROI region extraction in the scene, the seed pixel is selected with background subtraction, orderly, the neighborhood point is judged, the label of the primary selection seed is calibrated. The scene image segmentation algorithm is proposed based on snake model. The matching process is to find the maximum optimization process of the similar function, and the gradient drop method is adopted in mean shift algorithm. Extended kalman filtering is used to realize the robustness state estimation and prediction of the target tracking system. The results given by tracing experiment indicate that the proposed detailed algorithm is effective for partial loss of maneuvering target.

목차 (Table of Contents)

• Abstract
• 1. Introduction
• 2. FPGA Image Acquisition System
• 3. Wiener Filtering Algorithm
• 4. Region Growing Algorithm
• 5. Snake Model
• 6. The Mean Shift Iteration Algorithm
• 7. Extended Kalman Filter
• 8. Experiment and Analysis
• 9. Conclusions
• References