Research Of Rock Image Stitching Based On Sift Feature

Thin sections are an important research subject for reservoir microstructures. The form, range, grain size and fabric of the rock can be studied in detail under a microscope. However, with the limitation of the scope of microscope horizon, the overall characteristics and the local details of the rock thin section cannot be showed clearly. SIFT algorithm is a fast and effective algorithm for image feature extraction and the feature points extracted have good uniqueness and high compatibility.This article explains the process of image stitching technology, analyses the procedure of SIFT algorithm for extracting feature points. With an emphasis on the application of image stitching to rocks, the advantages and disadvantages of the SIFT algorithm is analyzed and improved. On one hand, the overlapping areas of the images are calculated and the feature points of the overlapping areas are extracted, so that matches can reduce the error rate. On the other hand, before matching the images, Principal Component Analysis(PCA) is used to reduce the dimension of SIFT feature vectors in overlapping area. Experimental results show that the new method reduces the computation of match, shortens the running time and improve accuracy.After the stitch of two images, this method is expanded to huge amounts of high-resolution microscopy images for stitching and a comprehensive super ultra-high resolution panoramic view of the micrograph is got. Because the existing computer software is difficult to fast display of large images, pyramid model is used to partition, hierarchy and storage and displaying hierarchically.Ultimately, researchers can observe the overview features under the large scale and study the local details under small scale.