Bulhakova T. Computer design of the built environment based on the modeling of visual perception

The dissertation is dedicated to the development of computer methods of compositional analysis and the formation of a built environment directly in the space of its three-dimensional model by means of geometric modeling of visual perception from selected points of view without using the perspective projections with quantitative determination of the characteristics of the visible environment. Three existing approaches (traditional, object, environmental) to compositional analysis of the built environment have been identified and analyzed in the paper. The third approach shows that the objectification of the features of human visual perception and the development of the methodology of quantitative determination of characteristics of built environment by means of computer technologies are available. The issue of the balanced development of the built environment regarding to visual perception of it by a human being is one of the pressing in modern architectural and design practices. It was proved that none of the existing methods of compositional analysis of the built environment from the position of visual perception cannot be chosen as an appropriate tool of architectural analysis. The reason of this statement is that the modern methods, which are based on the studying of perspective and orthogonal projections, have considerable limits while real and objective displaying of visual perception of environment by a human being. Thus, it was decided not to use the projections during the process of development of compositional analysis model of the built environment from the position of visual perception. We analyzed the characteristics of visual environment of its three-dimensional model based on modeling of visual perception by means of computer technologies directly in the area of the model without using perspective projections. It is offered to analyze the visual perception of any objects and their relations by means of using the solid angles with the vertices placed in the point of view and the surfaces that surround the visible contours of three-dimensional objects. This approach gives the opportunity to analyze the objects simultaneously regardless their position according to the observer; a part of that, the objects, which are accepted similarly in the reality, will have the same geometrical features during the modeling of visual perception and beside that, the refusal of using of the perspective projections will make possible to avoid the distortion of the images.