With the development of computer vision technology,panoramic image stitching has been widely used in fields such as scene reconstruction.A single traditional image cannot fully capture the panoramic view of the iconic...With the development of computer vision technology,panoramic image stitching has been widely used in fields such as scene reconstruction.A single traditional image cannot fully capture the panoramic view of the iconic East Gate of the South Campus of Shaanxi University of Technology.Therefore,this project aims to technically fuse multiple partial images into a complete panoramic image,enabling comprehensive recording and visual presentation of the architectural landscapes and spatial environments in this area.This report first introduces the technical background and application scenarios,clarifying the necessity of panoramic image stitching in campus landscape recording.It then elaborates on the core objectives and practical values,highlighting the role of technical solutions in improving image quality.Technically,a modular system design based on OpenCV is adopted,including modules such as image preprocessing,feature extraction and matching,image registration,fusion,and post-processing.Specifically,the SIFT algorithm is applied for feature extraction,KNN combined with ratio testing is used for feature matching,image registration is achieved by calculating the homography matrix,the fusion process utilizes multiband blending and Laplacian pyramid,and post-processing includes operations such as black area filling and CLAHE contrast enhancement.The experiment was conducted in a specific hardware and software environment using five overlapping images.After preprocessing,stitching,detail enhancement,and black edge repair,a panoramic image was successfully generated.The results show that the panoramic image fully presents the relevant scenery,with concealed seams,balanced exposure differences,and strong hierarchical details.This report provides a systematic description of the project’s technical implementation and achievement application.展开更多
文摘With the development of computer vision technology,panoramic image stitching has been widely used in fields such as scene reconstruction.A single traditional image cannot fully capture the panoramic view of the iconic East Gate of the South Campus of Shaanxi University of Technology.Therefore,this project aims to technically fuse multiple partial images into a complete panoramic image,enabling comprehensive recording and visual presentation of the architectural landscapes and spatial environments in this area.This report first introduces the technical background and application scenarios,clarifying the necessity of panoramic image stitching in campus landscape recording.It then elaborates on the core objectives and practical values,highlighting the role of technical solutions in improving image quality.Technically,a modular system design based on OpenCV is adopted,including modules such as image preprocessing,feature extraction and matching,image registration,fusion,and post-processing.Specifically,the SIFT algorithm is applied for feature extraction,KNN combined with ratio testing is used for feature matching,image registration is achieved by calculating the homography matrix,the fusion process utilizes multiband blending and Laplacian pyramid,and post-processing includes operations such as black area filling and CLAHE contrast enhancement.The experiment was conducted in a specific hardware and software environment using five overlapping images.After preprocessing,stitching,detail enhancement,and black edge repair,a panoramic image was successfully generated.The results show that the panoramic image fully presents the relevant scenery,with concealed seams,balanced exposure differences,and strong hierarchical details.This report provides a systematic description of the project’s technical implementation and achievement application.
