Simulation of Sunlight Environment and Optimization of Design Parameters in Urban Neighborhoods Based on Composite Adversarial Neural Network Generation Technology
摘要
This study investigates how urban block planning affects sunlight exposure. It aims to improve lighting conditions in high-latitude urban areas using scientific data. Current methods for designing roads and buildings in blocks rely on manual rules, which are laborious and struggle with complexity. The focus is on generating 3D urban models using machine learning, particularly Pix2Pix, to extract and learn from existing urban textures.