Automatic Understanding and Mapping of Regions in Cities Using Google Street View Images

José Carlos Rangel,Miguel Cazorla,Edmanuel Cruz

doi:10.3390/app12062971

Abstract

The use of semantic representations to achieve place understanding has been widely studied using indoor information. This kind of data can then be used for navigation, localization, and place identification using mobile devices. Nevertheless, applying this approach to outdoor data involves certain non-trivial procedures, such as gathering the information. This problem can be solved by using map APIs which allow images to be taken from the dataset captured to add to the map of a city. In this paper, we seek to leverage such APIs that collect images of city streets to generate a semantic representation of the city, built using a clustering algorithm and semantic descriptors. The main contribution of this work is to provide a new approach to generate a map with semantic information for each area of the city. The proposed method can automatically assign a semantic label for the cluster on the map. This method can be useful in smart cities and autonomous driving approaches due to the categorization of the zones in a city. The results show the robustness of the proposed pipeline and the advantages of using Google Street View images, semantic descriptors, and machine learning algorithms to generate semantic maps of outdoor places. These maps properly encode the zones existing in the selected city and are able to provide new zones between current ones.

Highlights

Thanks to technological and scientific advances, everyone with Internet access can find free and interactive maps of all kinds, such as Google Maps and applications built over them, such as Google Street View (GSV). The latter is used in research such as [1], where it contributes to the design and evaluation of a new tool for the semiautomatic detection of ramps in GSV images using computer vision and machine learning
We focus on generating a semantic representation of outdoor places using external tools and deep learning techniques
Due to the large number of maps produced by the combination of parameters, we decided to only consider the maps that produced a number of clusters inside a defined range, based on the possible number of zones that can occur in the selected city

Summary

Introduction

People have always needed to know about the place they live in and the area within which they move, and have represented the known space in writing or through drawings. Maps are scale representations of a territory, which, when developed with metric properties, help measure different parameters with great accuracy. A map can provide useful data for different activities, such as the analysis of urban plans. Thanks to technological and scientific advances, everyone with Internet access can find free and interactive maps of all kinds, such as Google Maps and applications built over them, such as Google Street View (GSV). The latter is used in research such as [1], where it contributes to the design and evaluation of a new tool for the semiautomatic detection of ramps in GSV images using computer vision and machine learning. GSV has been used as a tool for others works, such as [2–4], as its utility for scene-understanding approaches has been demonstrated

Methods

Results

Conclusion