Cylindrical Panoramic Image Stitching Based on SIFT Algorithm in Photogrammetry Systems

With the continuous development of photogrammetry technology, image stitching techniques have been widely applied in various fields, especially in 3D modeling, urban planning, and virtual reality. Cylindrical panoramic images, as a common measurement method, are often used to capture complete views of environments with high precision. However, the process of image stitching is challenged by issues such as image distortion and difficulty in matching overlapping regions, which can affect the stitching results. To address these challenges, this paper proposes a cylindrical panoramic image stitching method based on the SIFT (Scale-Invariant Feature Transform) algorithm. The method first extracts feature points from the images using SIFT, and then utilizes feature point matching and geometric transformation techniques to achieve precise image stitching. Experimental results show that the proposed method significantly improves the accuracy and stability of cylindrical panoramic image stitching. This research provides new ideas and technical support for automatic image stitching of cylindrical panoramic images in photogrammetry systems.