Environmental Landscape Art Design Based on Visual Neural Network Model in Rural Construction

Abstract As the resources of social development continue to tilt to the countryside, the speed of rural construction continues to accelerate. In recent years, because of the higher quality of lifestyles, the demand of rural environment landscape art has gradually increased. In order to assist rural construction and improve the artistic quality of its environmental landscape, this paper proposes an environmental landscape art design method based on a visual neural network model. Firstly, the Swin Transformer text encoder is used to characterise the landscape art demand in rural construction. Then, the text feature vector of landscape art demand is input into the GAN model to generate the image content of rural construction. Finally, to better evaluate the landscape art level of the above methods in this paper, we propose an evaluating method for the landscape designing tasks. We conduct the experiments and achieve the FID value of 15.23, which can demonstrate that our method can effectively carry out an environmental landscape design for rural construction and simplify the process of rural construction. The landscape design evaluation method can evaluate the environmental landscape design accurately by the accuracy of over 80 %, and further improve and optimise the acceptance link of rural construction.