Research on Landmark Matching for an Unmanned Aerial Vehicle

A mathematic model for landmark matching is proposed based on pattern recognition methods. The relevant aspects that influence the performance of matching algorithm are analyzed. Normalized correlative reference image matching is employed in landmark matching. After sufficient simulation experiments, the following conclusion is made: The algorithm can meet the requirements for landmark positioning at a low signal to noise ratio. In addition, the genetic algorithm is used in the image-matching search. The analysis of the computational complexity in the experiments is presented. The results prove that this method can reduce the computation load greatly. Thus, the real time ability is improved.