Research And Application Of 3D Modeling For Generalized Cylindrical Object Based On Single Image

With the development of computer and mobile terminal hardware,3D modeling based on images has become a hot topic in the field of computer vision and computer graphics.On the one hand,3D model has a wide range of applications such as rapid product design,product exhibition,3D printing,simulation,visualization,game animation,augmented reality and virtual reality,which offers a strong visual experience,wonderful interactive and immersive experience.It also plays a positive role in protecting the cultural heritage and cultural heritage digitization.On the other hand,we live in a world with a large number of images,and it is easy to capture images and process them.Based on this background,we present an interactive 3D modeling method based on single image.In this thesis,the reconstructed object using our method should be a kind of generalized cylindrical object.The main idea is to preprocess the input image with some simple interactions by the user,including the detection of feature points,image segmentation,symmetry axis,body contour and component contour.Then,we propose four structural elements to define the structural features of the object in the image,including Base,Layer,Cap and Side.Four structural elements can be created with the user operations.To model the non-symmetric object,the morphing curves are generated.The morphing curves among adjacent layers make it possible to generate a number of inter layer structures between adjacent layers to complete the smooth,transitional and irregular surface modeling task.For the symmetric object,the adjacent layers are firstly constructed with translation,rotation and scaling transformation,then the geometric relationship between adjacent layers is calculated.These geometric constraints are transformed into three dimensional information such as points,lines and surfaces.Finally,3D mesh is created and mapped with the texture.Multiple formats of the 3D mesh is supported.The experimental results show that the proposed method can generate 3d models and can be used in many fields such as rapid product design,3D printing,cultural relics digitization,virtual display and augmented reality.