Research And Application Of Full-length X-ray Image Stitching Algorithm For Orthopedics

When performing orthopedic surgery on the human spine or other parts,sometimes it is necessary to obtain an X-ray anatomical structure map of the complete object,such as the full length of the spine,to assist the doctor in evaluating the operation and the quality of the operation.At present,due to the limitation of factors such as size and shooting position of the X-ray imaging device,the field of view of the acquired X-ray image is often narrow,and the entire lesion area cannot be completely imaged by one filming.At present,the mainstream method is to use X-ray imaging equipment to take multiple shots by region,obtain segmented medical X-ray images about the human spine and other orthopedic parts,and then form a complete panoramic image through postprocessing of the image workstation.This method requires specific the equipment is timeconsuming,and it is difficult to meet the evaluation requirements of X-ray images of the whole area during the operation.With the widespread use of digital image processing technology in medical imaging,digital medical image stitching can achieve orthopedic X-ray images to be quickly and accurately stitched in the operating room,providing a reliable basis for doctors' intraoperative evaluation.This article mainly implements a stitching technology for medical orthopedics X-ray images.The main contents are as follows:In the determination of the basic process of medical X-ray image stitching technology,after summarizing the research results in the field of medical image stitching,this thesis determines the acquisition and pre-processing methods,image registration methods and image fusion methods of medical X-ray images.In the feature extraction stage,this thesis aims at the problem that the SIFT algorithm detects fewer feature points on medical orthopedics X-ray images in practical applications,and the difference in the number of feature points detected between different medical orthopedics X-ray images The big problem is to propose an improved SIFT algorithm that automatically adjusts the contrast threshold based on image contrast information.It mainly adjusts the contrast threshold in the SIFT algorithm by defining the image contrast value,so that the number of feature points extracted on different medical orthopedic Xray images Both are significantly improved and are in a suitable range.In the feature matching stage,this thesis proposes to use the Pearson correlation coefficient to further match and filter the repeated matching point pairs to eliminate the "one-to-many" or "many-to-one" problems between matching point pairs after using Euclidean distance matching.The redundant and repeated matching point pairs ensure the unique matching relationship between the matching point pairs and improve the matching accuracy.The research is based on the platform of multiple medical orthopedics X-ray image stitching system.The stitching system is used to automatically stitch medical orthopedic X-ray images taken by X-ray imaging equipment.Experiments show that the target in the complete X-ray image after stitching is complete and the details are clear.Compared with the traditional medical image stitching method,this method can reduce the running time while ensuring the accuracy of medical X-ray images with high noise and low contrast,and has certain application value for intraoperative evaluation of clinical orthopedic surgery.