Research On Key Technology Of Three-Dimensional Model Retrieval Based On Ontology

With the rapid development of computer science and technology, realistic effect when the computer display three-dimensional model, promoting the three-dimensional model of the widely used, those makes the rapid growth of three-dimensional in recent years. However, how to effectively manage the vast amounts of three-dimensional model are currently the main problems. The traditional approach is to complete the three-dimensional model for manual classification and database built on the participation of expert in field, so that spent a lot of manpower and resources. Three-dimensional model of the underlying characteristics or labeling text words were used to retrieve, but the retrieval result is not good.For the above problems, this paper presents a semantic-based three-dimensional model of management method, first combination Hownet perfect ontology, then improved ontology method is applied to ontology retrieval, last complete the three-dimensional model retrieval based on ontology. The main work of this article:(1)Realization of a combination of Hownet extended ontology method. By analyzing the original ontology, get the basic concept of concept in original ontology use the Hownet. Basic concept add in the ontology improve the ontology concept; And use Hownet calculate the concept of similarity in the ontology, the relationship between the concept to join the ontology, complete the expansion of the ontology concept relations.(2)Semiautomatic semantic annotation method is applied to the ontology construction of three-dimensional model. On the basis of typical model of a given semantic annotation, compare the Similarity between the typical model and sign model, and automatic semantic annotation of a given threshold model.(3)Improvement in the general semantic retrieval, two kinds of processing is considered, in the three-dimensional semantic retrieval:①hen the retrieval terms have some described in the OWL documents and the system for semantic retrieval based on user search terms.②When the retrieval terms have not some described in the OWL documents, calculation the similar of two words provided by Hownet, one word is the keywords and the other word is the conception, object or property in the ontology library, and put the most similar words, as the search terms entered by the user semantic retrieval.Designed a three-dimensional model based on ontology management system in this paper. It completed the ontology construction and ontology reasoning and retrieval, the retrieval methods than traditional retrieval methods in recall and precision has been a significant improvement, the general ontology retrieval result for empty issues is resolved.