Identification of trip generators using remote sensing and geographic information system

Abstract

The traditional methods of building and updating databases for transport analysis that address origin/destination (O/D) data collection, such as surveys and traffic counting, are generally manual, costly and time-consuming. This serves as an impediment, especially in developing countries, owing to factors such as scarcity of financial resources, lack of skilled personnel, and high rate of urban criminality that dissuades interviewers from making visits and threatens equipment integrity. In this paper, we propose an alternative methodology that is more agile, cheaper, and safer in comparison, which identifies zones with clusters of trip generators (TG), i.e., areas with the highest potential to generate trips. This method utilizes the object-based image analysis approach to extract from satellite images, zones with high potential to generate trips in the city of João Pessoa in Brazil. It aims to map the land use/land cover (LULC) changes through image classification and associate the resulting LULC classes with the urban trip generation (O/D) data. By identifying zones that generate the most number of trips, the results can help to devise guidelines for conducting fieldwork that can facilitate the creation of a comprehensive repository to improve transportation planning. The identification of TG can supplement or provide preliminary measures for scouting prospective O/D zones. This paper constitutes an effort to integrate remote sensing into the urban planning process, primarily in areas where readily accessible data on transportation facilities and performance either does not exist or is rendered obsolete, which is often the norm in developing countries.