Research On View-Based Retrieval And Edit Method For 3D Tree Model

As an important resource of domain-specific 3D model, tree models have being widely used across lots of areas. However, the problem that modeling process for 3D tree model is too complex has never been thoroughly resolved. To avoid the difficulty in tree modeling, reusing existing models resource become an effective way to meet the demand of using tree model. With the enrichment of the tree model resources, reuse of 3D tree models will not only play an important value, but also face more difficult and challenge. Querying the target model among various shape of 3D tree models will getting harder, and accurate retrieval result can’t be achieved with existing methods. There is also lack of simple and efficient edit method for 3D tree model, which will make the reuse process be troubled by complex operations and unsuccessful results.To solve the two key issues of tree models reuses, this thesis carried out research on view-based retrieval and edit method for 3D tree model, and aim to achieve intuitive and easy-to-use framework for reusing 3D tree model resource. First, this thesis proposes a progressive retrieval method based on morphology of 3D tree model, to gain accurate retrieval result in tree models resource; second, this thesis introduce an adaptive editing method guided by view description, to lightens the workload of re-editing process, and make reuse process becomes more convenient and fast.The primarily work and achievements of this thesis is as follow:(1) In order to represent tree object in a simple and efficient form, this thesis adopts hierarchical local view to describe 3D tree model. This form of representative view not only unifies the retrieve and editing intentions for model-user, but also highlights the important features of 3D tree model, which eventually help to implement and link all key parts of reuse process.(2) In order to improve retrieval accuracy and efficiency for 3D tree models, this thesis design a progressive retrieval method based on local view of tree model. This method using representative view to express search intentions and key feature of tree model, and extract view features with its corresponding semantic information, then adopted view feature into a multi-stage retrieval process that achieves a coarse-to-fine retrieval performance for 3D tree models(3) In order to reduce workload of editing 3d tree models, this thesis presents an adaptive editing method guided by view description. This method associate representative view with its corresponding part of tree model, then guides tree model into transformation stage by view information. Meanwhile, this method adjust shape and structure of tree model based on botanical rules after editing stage, make sure the tree model have realistic-looking after adaptive editing.(4) In order to validate the effectiveness of progress retrieve and adaptive editing method, this thesis design and implement 3d tree model retrieval and edit system. This system collect and integrate existing tree model resources to building 3D model library, and providing user-friendly interface to make retrieve and editing process more intuitively and efficiently, ultimately make 3D tree model resource can be underused.