Development of a City-Scale Approach for Façade Color Measurement with Building Functional Classification Using Deep Learning and Street View Images

Precise measuring of urban façade color is necessary for planning. The existing manual methods building are limited by time and labor costs hardly carried out on a city scale. These also make it challenging to identify the role function in controlling guiding This paper explores city-scale approach measurement with functional classification using state-of-the-art deep learning techniques street view images. Firstly, we used semantic segmentation extract façades conducted calibration photos pre-processing collected Then, proposed chart-based method multi-label learning-based method. Next, field survey data were as ground truth verify accuracy classification. Finally, applied our generate distribution maps three metropolises China, results proved transferability effectiveness scheme. can provide managers an overall perception across areas cost-efficient way, contributing data-driven decision making analytics