Research on Feature Extraction and Match Method based on the Surf Algorithm for Mobile Augmented Reality System

Abstract

According to the characteristics of constrained mobile devices, an improved Surf feature extraction and matching algorithm is proposed. First, dynamic pyramid is established in order to meet the real-time requirements; then reducing redundant feature points and strengthening the feature points of big difference individuality is done based on the Niche ideas; finally improved classification BBF matching algorithm is used to speed the matching. The contrast experiment on Surf and its improved algorithm is done based on the iOS platform. The experimental results demonstrate that the improved Surf algorithm has strong robustness, fast speed, high accuracy, and lays the foundation for mobile augmented reality technology in video, animation, text, 3D virtual scene registration, so it has great application value.