Image registration method of Large Volume Parenteral based on improved ORB algorithm

Aiming at the problem of vibration interference caused by mechanical vibration in the intelligent detection of Large Volume Parenteral, this paper studies an improved image registration algorithm that based on the ORB algorithm to compensate for the image position deviation. Firstly, the median filtering method based on extreme value detection is used to remove the interference of random noise; then the image enhancement link based on the Laplacian operator is added for the improvement of the characteristics of dark images; and the Brief descriptor is improved in the feature point description stage, using The $3 \times 3$ neighborhood sampling method is used to replace the original single-point sampling method to avoid noise point interference; finally, the block matching strategy is introduced in the feature point matching stage to improve the matching efficiency. The experimental comparison shows that the improved registration algorithm has better registration effect while maintaining the fastness.