Simulation Study on Extraction Method of Graphic Elements of Ocean Plane Based on Visual Communication

ABSTRACT Liu, M.; Wang, J., and Zhou, J., 2020. Simulation study on extraction method of graphic elements of ocean plane based on visual communication. In: Yang, D.F. and Wang, H. (eds.), Recent Advances in Marine Geology and Environmental Oceanography. Journal of Coastal Research, Special Issue No. 108, pp. 113–117. Coconut Creek (Florida), ISSN 0749-0208. One of the principles of ocean architecture design is to reasonably coordinate the use of architectural plane graphics to match the structure. In addition, from the perspective of the development of architectural history, the symbols related to visual communication, such as graphics and patterns, are applied to the building skin and bear specific information and meaning, which has occurred throughout the development of architecture. With the deepening of interdisciplinary degree, visual elements and visual symbols are reinterpreted by combining visual communication with the design of marine architecture. The collection of marine architectural plane elements based on visual communication plays an important role in the realization of marine architectural design. Traditional methods tend to focus on the extraction of attributes; the accuracy is low and the recognition rate is not high. This method first marks the extracted graphic elements, separates the background and the main elements, preprocesses the graphic elements, and then obtains the eigenvalues of the graphic elements by using the covariance matrix. After the average calculation of the eigenvalues, the variance of the eigenvalues of the elements with high dimensions is reduced. The grayscale map is used to determine the size and direction of the feature points of the core elements, and the final recomposition is carried out to realize the extraction of graphic elements of the ocean building plane under visual communication. The simulation results show that the graphic elements extracted by this method are not only accurate but also of low dimension.