Improved Feature Extracting and Matching Algorithm Based on FAST and Local Binary Pattern

This study introduces an enhanced version of the FAST algorithm aimed at achieving a uniform distribution of feature points. We also developed a feature descriptor based on local binary patterns (LBP), which provides rotation invariance and robust anti-noise capabilities. Furthermore, we have designed specific feature-matching algorithms tailored for these descriptors and created methods to identify and eliminate mismatched features. Finally, we thoroughly assess the algorithm's real-time performance and locating accuracy using a publicly available dataset. The experimental results demonstrate that our algorithm performs well in localization accuracy, with a running time that is approximately 10% of that of the SURF algorithm and just 1% of that required by the SIFT algorithm.