Research On 3D Model Semantic Auto-annotation Based On Ontology

With the rapid evolution of computer software with hardware and computer theories, 3D model has been widely used in virtual reality, mechanical manufacturing and other industries as a main data type, specially in 3D model semantic annotation as well as semantic retrieval applications. however, Semantic knowledge of a product such as product usage, functional characteristics and design principles is essential during its conceptual design stage to guide the design process. however, content-based 3D model retrieval focuses on product geometric similarity rather than their functional similarity. 3D model retrieval based on functional semantics is more significant than content-based 3D model retrieval during the mechanical conceptual design phase. Therefore, a functional semantic-based 3D model annotation is a work which has theoretical, economic value.A functional semantic-based 3D model auto-annotation has become one of the hotpot researches, and it is also an important part of computer visualization, pattern identification and 3D data processing. In this paper,a functional semantic-based 3D model annotation are detailed described and summarized, and an novel 3D model annotation method is put forward. The major work of this paper are as follows:1. We illustrated the background and significance of functional semantic-based 3D model auto-annotation method research. the 3D model functional semantic ontology is constructed to formally represent the functional semantics of 3D model and describe the mechanical conceptual design space comprehensively and consistently. constructing 3D Model Repository that supports model auto-annotation based on functional semantics. inspire designer creativity through existing STEP model, shorten design cycle, and reduce costs.2. When extracted the shape feature, the existing retrieval method only calculated the whole information of 3D model, and ignored the partial information of 3D model. By combining and improving existing algorithms, an approach to represent STEP model as attributed adjacency graphs is proposed. First, In this method, the geometry and topology data are extracted from STEP model. Then By extracting data from a STEP file layer by layer, a STEP model is represented by using an AAG, in which graph nodes correspond to STEP model surfaces and graph links correspond to STEP model edges. Last, On the basis of AAG, the functional semantics of STEP model are annotated semi-automatically by matching STEP model that contain the partial features of which functional semantics have been annotated manually.3. In case one function may be not a comprehensive description of the target model, a 3D model retrieval algorithm that supports function extended retrieval is proposed, which is used to calculate function semantic similarities between models to explore more potential creative design knowledge in the semantic level. The proposed research addresses actual needs and presents a new way to acquire STEP model in the mechanical conceptual design phase. an example is taken to illustrate that the search result can satisfy the requirements of conceptual design and contribute to design reuse and innovative design.