Fusing Taxi Trajectories and RS Images to Build Road Map via DCNN

Yali Li,Caili Zhang,Longgang Xiang,Huayi Wu

doi:10.1109/access.2019.2951730

Abstract

Accurate road maps are fundamental to a wide range of applications, such as navigation, transportation and urban planning. With the rapid development of Globe Position System (GPS) and Remote Sensing (RS), abundant and continuously updated spatiotemporal data are available for road map building. However, using single data source inevitably results in limited performance on road information recognition, due to inherent defects in data, such as the noise, low sampling rate and uneven density distribution in trajectories, and occluded roads in RS-images. This paper aims to integrate GPS trajectories and RS-images to build road map, which is composed of intersections and links. First, both dynamic and static trajectory characteristics are utilized to capture road features indicated in trajectories, and meanwhile, road features implied in by RS-images are extracted via transfer-learning with an advanced deep neural network. Then, the two distinct types of road features are explored together to learn the potential area of road centerlines based on a variation of U-Net, which refines the U-Net to incorporate the structure characters of road. Finally, a multistep refinement is introduced to extract road centerlines through morphology processing, then derive intersections through connectivity analysis and detect links based on sliding-window. In addition, inspired by vector rasterization, a fast and automatic strategy is developed to create large-scale road dataset to significantly reduce labor and time. Compared with three state-of-the-art segment networks, the proposed network achieves the highest correctness and quality. The experimental results demonstrate that trajectory data and RS image complement each other, which assure the extracted road network possesses both integrity and connectivity.

Highlights

Road network is a fundamental element in both geographic information system (GIS) and intelligent transportation system (ITS) [1] and plays an essential role in the automatic vehicle navigation, traffic management and urban planning [2]
Xu et al [1] developed a segmentation model based on densely connected convolutional networks (DenseNet) considering local and global attention to extract road from remote sensing images, and the results demonstrated better harmonic mean accuracy than some machine learning methods
The data of these three areas are split into two subsets: one training set including two regions, marked with a blue border, and one validation set based on the rest one region, marked with a red border

Summary

Introduction

Road network is a fundamental element in both geographic information system (GIS) and intelligent transportation system (ITS) [1] and plays an essential role in the automatic vehicle navigation, traffic management and urban planning [2]. It is necessary to provide accurate and up-todate road digital map. The road structure is changing rapidly due to rapid development of cities. Conventional methods to build and update urban road maps are timeconsuming, labor-intensive and costly [3]. With the advance in global positioning system (GPS) and remote sensing (RS), a large amount of spatiotemporal data containing road information become available, making the automatic, timely and efficient road map production possible.

Objectives

Results

Conclusion