Multi-Feature Fusion Based on Particle Filter Algorithm for Moving Target Tracking

Abstract

Robust moving target tracking has become an important topic in the field of computer vision. The fusion of feature such as color, texture, edge strength and motion has proved to be a promising approach to robust visual tracking in situations where no single feature is suitable. Due to the poor tracking performance in complicated scenarios, a new visual target tracking scheme is proposed. This algorithm exploits the fusion feature of color feature and local binary pattern (LBP) texture feature under the framework of particle filter to improve the robustness of target tracking. In order to get the accurate color model of target, the multi-part color histogram with spatial information is introduced. Comparing with the particle filter tracking algorithm based on single feature, experimental results show that the proposed method is more robust in terms of pose changes and illumination changes, especially in scenarios where the target object contains cluttered background with similar color distributions.

Keywords

• Target tracking
• Particle filtering
• Multi-part color representation
• LBP texture features.