The Format Criterion Of Master Thesis

Image-based model reconstruction technique is the use of the geometric information of the image itself to restore the three-dimensional model of the corresponding. Relative to the traditional way, Image-based model reconstruction method have fast, short circle, realistic and so on characteristics. With the rapid development of computer technology, model-based image reconstruction technique has drawn increasing concern of the researchers. Image-based building model reconstruction techniques has been important research topic in computer vision field, it has a wide application in the digital city, restore ancient buildings, virtual reality and other fields.The article’s background is virtual campus design and application projects, reconstruct the building model in the virtual campus, analysis deeply the current building model reconstruction based on image, study deeply the key technologies of the three-dimensional reconstruction, through the camera calibration, feature point markers, determine the base plane, texture mapping and other key technologies, get the reconstruction of building models based on a single image in the end. The paper accomplishes main work and researches are as follows:1) Camera calibration method based on geometry constraints. For the building have parallel, orthogonal and so on geometric characteristics, I propose a camera self-calibration algorithm on geometric. Draw three sets of parallel lines through human interaction in the perpendicular direction, establish the linear equations, and convert into a coefficient matrix, and use the least squares method to calculate the vanishing point coordinates, then determine the camera parameters, completed the camera calibration.2) Building model reconstruction method based on a single image. For certain scene, especially the scene of man-made buildings, I propose a building model reconstruction method based on a single image. Determine the two-dimensional space and three-dimensional perspective projection relations through the camera calibration recover the three-dimensional space point through the linear camera model; determine the base plane, and according to the geometrical relationship of the other plane and the base plane, recursive restore all planes in the building, and then complete the reconstruction of the model.3) Building model texture mapping method. Analysis the OpenGL texture mapping mechanism, with the singularity of the building surface texture, I propose a building model for texture mapping method. Extract the texture pattern by image processing tools, calculate the image feature points corresponding to the texture coordinate, and map the corresponding texture to the reconstruction of three-dimensional model, and then gain the realistic3D model.4) Design and implementation of the building model reconstruction system based on image. According to the actual needs of the system design, design system framework, resolve the key issues in the implementation of the system, achieve and apply the camera calibration method based on geometry constraints, building model reconstruction method based on a single image and building model texture mapping method and other achievement.