Urban Safety Perception in the Age of Big Data: Modeling Residents’ Perceived with Street View Imagery, Deep Learning, and Spatial Simulation

Leveraging massive Baidu Street View datasets, convolutional neural networks, and spatial regression techniques, this book uncovers how visual features—such as greenery, lighting, cleanliness, and built structure—affect safety perception at fine spatial scales. This book presents an innovative spatial framework for understanding and simulating urban residents’ safety perception using street-level imagery and deep learning. Bridging GIScience, environmental criminology, and urban informatics, it explores how micro-scale urban environments shape fear of crime and perceived safety, offering both theoretical insights and practical modeling tools. Leveraging massive Baidu Street View datasets, convolutional neural networks, and spatial regression techniques, this book uncovers how visual features—such as greenery, lighting, cleanliness, and built structure—affect safety perception at fine spatial scales. It further integrates survey data, crime records, and machine learning to simulate perceived safety across neighborhoods. Designed for researchers and professionals in GIS, urban planning, public health, environmental psychology, and smart city development, this book is suitable for advanced students and interdisciplinary scholars seeking new methods in spatial perception modeling.