Study Of Image Retrieval Based On Fusion Of Color And Texture Features

The CAD technology in architecture currently plays an auxiliary role in designing and drawing in architecture designing and construction. Paper construction drawings are still the only medium from designing to drawings inspection, budgeting and the construction. In each stage, drawings can be only read by human with slow efficiency, frequent mistakes and unevaluated accuracy affected by human's objectivity and subjectivity. The main reason for this is that CAD data can not be shared among designing, drawings inspection, budgeting and construction. Thus, the graphics recognition of CAD documents and the extraction of project quantity information will be of research and practical value both at present and in the long term. Its future application is promising.This paper, with architecture designing and the project budgeting integration as the background, studies the the graphics recognition based project information obtaining technology. This paper firstly elaborates the architectural components with semantic analysis according to the composition and features of the construction drawings and introduces the technology of CAD secondary development. Then, the paper discusses the ideas and methods of the graphics recognition and studies the search strategies of graphics and graphic elements. The symbol recognition based on statistics and structures is an important method in the pattern recognition field. Moreover, the paper analyzes the representation and recognition features of architecture symbols and studies the recognition principle, sequence and methods of architecture symbols based on the feature matching. Taking wall symbols for example, the paper analyzes partial elements of the wall symbols, presents the recognition steps of the wall symbols, studies the extraction algorithm of wall symbols and finally comes to the inference rules of the wall symbols. On this basis, the whole building including doors, windows and rooms can be recognized.Finally, the paper summerizes the research on the project information obtaining technology based on the graphic recognition and it also presents future research.