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

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. INDEX TERMS Road map generation, crowdsourcing trajectory data, remote sensing image, data fusion.

A Guided Deep Learning Approach for Joint Road Extraction and Intersection Detection From RS Images and Taxi Trajectories

“…Therefore, it is necessary to integrate RS and trajectories in road information extraction and find an efficient fusion method to combine their respective advantages and improve the quality of road extraction. Although there are some preliminary studies on road extraction based on fused RS images and trajectories, the technology of data fusion is relatively simple, such as the concatenation of two data sources directly as the input of the deep learning network [14,15]. The lasted fusion method is DeepDualMapper [16], which designed a gated module to learn a complementary-style fusion.…”

“…However, to a certain extent, road extraction and intersection detection depend on each other. The road detection result contains intersections, since it is easy to obtain intersections from road detection by using tensor voting [19] or by analyzing the connectivity of the morphological thinning result of the road [15]. Meanwhile, the intersections can help to link broken roads around junctions.…”

“…One method is to extract roads from one data source based on the processing results of the other data source [41]. The other methods concatenate RS images and GPS trajectory feature maps as the inputs of deep neural networks [14,15]. These fusion methods are not sufficiently delicate to fully mine the enhanced information between two data sources, nor do they consider obtaining various road information simultaneously based on fusion data.…”

Generating Road Networks for Old Downtown Areas Based on Crowd-Sourced Vehicle Trajectories

“…How to obtain road information for reasonable planning and resource allocation has always been an economic issue for national economies and people’s livelihoods [ 1 ]. With the development of surveying, mapping, communications, computers, and other technologies, we can infer road networks based on various data sources, such as crowd-sourced vehicle trajectories [ 2 , 3 , 4 , 5 ], laser point clouds [ 6 , 7 ], remote sensing images [ 8 , 9 ], aerial images [ 10 , 11 , 12 ], OpenStreetMap [ 13 , 14 , 15 ], etc. Among these data sources, crowd-sourced trajectories have become mainstream data sources of generating road information, and have triggered a large amount of research on road extraction in the past few years, focusing on prominent features, such as wide coverage, high update frequency, and low acquisition cost [ 16 ].…”

Aerial Images Meet Crowdsourced Trajectories: A New Approach to Robust Road Extraction

“…It is very natural to incorporate aerial images and crowdsourced trajectories to extract traffic roads robustly. However, there are very limited works [28], [29] that simultaneously utilized the two modalities mentioned above. Moreover, these works directly fed the concatenation of aerial images and rendered maps of trajectories or their features into convolutional neural networks, which is a suboptimal strategy for multimodal fusion [30].…”