Urban virtual environment landscape design and system based on PSO-BP neural network

Abstract

AbstractIn the last few years, with the fast growing of neural network field such as those for virtual reception and enhanced nature, the practice and theory of conventional landscape are impacted and challenged by virtual landscape based on these sorts of neural network technologies. On the one hand, the virtual landscape changes the carrier of landscape design from material real world to the networked virtual world, which breaks the traditional way of generating landscape and the way of expression of results. On the other hand, the virtualized and networked morphological characteristics of the virtual landscape itself and its capacity that can offer users a sense of immertion, interplay and enjoyment of the experience provide a way of extending and deepening the realm of scenery. It is also a new type of landscape that conforms to the trend of the times created in the background of the fast evolution of scientific and technical development. Virtual landscape brings new construction thinking and practical means for the application of digital city, the construction of urban context, and the development and utilization of relics. It provides an important research source for thinking about the relationship between current humanities and science, material and virtual, history and contemporary. After the research and experiment on the urban environment landscape design of PSO-BP neural network, the experimental data showed that before using the neural network method to improve Yanta East Garden, 57% of the crowd were satisfied with the overall impression of Yanta East Garden, and 17% were dissatisfied. After the improvement, 67% were satisfied with the landscape of Yanta East Garden, only 5% were dissatisfied, and the landscape satisfaction increased by 10%. The survey group believed that the landscape color of Yanta East Garden was full of historical flavor, especially the small sculptures convey the unique Qin Opera culture. The above data show that the method based on neural network is very suitable for the improvement and development of urban landscape design.