Anti-fake Legend Registration Algorithm Based on Improved ORB

For the anti-counterfeiting code, there are interference factors such as character adhesion, strong background, complex lighting environment, etc., which lead to the failure to correctly recognize each character, propose a legend anti-counterfeiting method that takes a specific area containing the anti-counterfeiting code as the processing object, and then designs an image registration algorithm based on improved ORB (oriented FAST and rotated BRIEF) for this anti-counterfeiting method. First, use the adaptive quadtree algorithm to improve the FAST algorithm to extract evenly distributed feature points;then use the rBRIEF algorithm to describe the feature points, and the KNN algorithm to roughly match the feature points;finally, an improved random sampling consensus (RANSAC) algorithm is used,a TL (Triangle-like) algorithm is proposed, which obtains a new sample set by judging whether two triangles composed of three pairs of feature points are similar, and calculates the best transformation matrix to eliminate mismatched points. The experimental results show that the improved algorithm has strong real-time performance, accuracy and anti-interference ability in anti-counterfeiting legend, and the practical application effect is stronger.