Multisource Fusion of Remote Sensing and Mobility Data for Traffic Analysis Zone Definition

The spatial discretization into traffic analysis zones (TAZs) is a crucial yet often overlooked step in transport systems analysis in an urban area. Beyond general guidelines and case-specific experience, the formulation of an automatic or semiautomatic approach remains an open challenge. Nevertheless, the joint availability of remote sensing images, whose segmentation allows identifying homogeneous regions and separating built and non-built areas, of databases that collect information on the activities performed therein (shops, workplaces, etc.), and of floating car data (FCD) provided by GPS sensors on mobile phones and cars, exhibits a remarkable potential in this framework. This paper introduces a novel method for the automatic definition of traffic analysis zones in urban areas, based on the multimodal fusion of satellite multispectral imagery and ancillary geospatial data (OpenStreetMap, land cover, and FCD) and aimed at maximizing homogeneity within zones and minimizing the number of trips with origin and destination in the same zone. The proposed approach integrates the above data sources to generate a TAZ configuration that can be directly applied in travel demand forecasting models. The methodology was applied to a case study in a neighborhood of Rome, Italy.