Research on Automation of Rural Landscape Design Based on Convolutional Neural Network and Generative Adversarial Network

Driven by the rural revitalization strategy, rural landscape design is entering a new stage. It needs to face the dual challenges of innovation and sustainability. This study explores the application of deep learning in rural landscape design and analyzes its potential in constructing design patterns and optimizing design concepts. The natural environment, cultural heritage and historical evolution data of the countryside were collected. The data were preprocessed and the CNN and GAN technologies were used to construct an automatic generation model for landscape design. The results of the experiment showed that the quality of GAN image generation exceeded that of the traditional CNN model. The output images were clearer, more detailed, and more similar in structure, which more accurately reflected the individual characteristics of the rural scenery.