FuNet: A Novel Road Extraction Network with Fusion of Location Data and Remote Sensing Imagery

Zhou, Kai; Xie, Yan; Gao, Zhan; Miao, Fang; Zhang, Lei

doi:10.3390/ijgi10010039

Cited by 24 publications

(11 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the extraction of roadways from a complex urban landscape, the integration of border and contextual information plays a significant role. Zhou et al, [18] explored the image pre-processing technique with histogram equalisation resulting in better contrast enhancement by nearly one per cent. Due to limited resources, the authors combined the prediction output with the original image.…”

Section: Literature Reviewmentioning

confidence: 99%

Road segmentation: exploiting the efficiency of skip connections for efficient semantic segmentation

Madhumita

Hariharan

Devendra

et al. 2023

J. S. Asian Logistics Transport

View full text Add to dashboard Cite

Extraction of features is a unique approach that has several real-time applications. In remote sensing, road feature extraction includes recognising and extracting essential information about roads from high-resolution remote sensing photos. This data may include the shape, size, direction, and connection of roadways and their surroundings we can increase the accuracy and efficiency of road recognition algorithms and better comprehend the geographical distribution of roads in metropolitan settings. Nevertheless, traditional techniques of road recognition sometimes rely on manual labour or hand-crafted characteristics, which may be costly and time-consuming. Convolutional neural networks (CNNs) have been frequently employed for road detection due to their capacity to automatically learn characteristics from data. In this research, we provide Modified LinkNet for the extraction of road features from high-resolution remote-sensing images. The proposed model incorporates several variables resulting in superior performance compared to current cutting-edge road detection techniques. SpaceNet road detection dataset yields 0.81 mIoU, which is approximately 0.1, 0.11, and 0.03 mIoU greater than DeeplabV3+, and LinkNet, respectively (base architecture). In addition, the training of the model enables more efficient inference making our method applicable to real-time applications and are therefore a more effective and efficient solution for the extraction of road features from remote sensing images.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Road segmentation: exploiting the efficiency of skip connections for efficient semantic segmentation

Madhumita

Hariharan

Devendra

et al. 2023

J. S. Asian Logistics Transport

View full text Add to dashboard Cite

show abstract

“…Sun et al [24] added crowdsourced global positioning system (GPS) data to satellite images to extract roads with CNN-based semantic segmentation. Zhou et al [30] fused remote sensing images and GPS for road detection and extraction. Additional recent learning-based studies included Lu et al [16] proposing a multi-scale residual network for road detection, Pan et al [21] proposing a fully convolutional neural network using VHR remote sensing, Mattyus et al [17] estimated road topology from aerial images, Mi et al [18] generated road lane graphs from LiDAR data with a hierarchical graph generation model.…”

Section: Street and Road Network Map Generationmentioning

confidence: 99%

APE: An Open and Shared Annotated Dataset for Learning Urban Pedestrian Path Networks

Zhang¹,

Bolten²,

Mehta³

et al. 2023

Preprint

View full text Add to dashboard Cite

Inferring the full transportation network, including sidewalks and cycleways, is crucial for many automated systems, including autonomous driving, multi-modal navigation, trip planning, mobility simulations, and freight management. Many transportation decisions can be informed based on an accurate pedestrian network, its interactions, and connectivity with the road networks of other modes of travel. A connected pedestrian path network is vital to transportation activities, as sidewalks and crossings connect pedestrians to other modes of transportation. However, information about these paths' location and connectivity is often missing or inaccurate in city planning systems and wayfinding applications, causing severe information gaps and errors for planners and pedestrians. This work begins to address this problem at scale by introducing a novel dataset of aerial satellite imagery, street map imagery, and rasterized annotations of sidewalks, crossings, and corner bulbs in urban cities. The dataset spans 2, 700𝑘𝑚 2 land area, covering select regions from 6 different cities. It can be used for various learning tasks related to segmenting and understanding pedestrian environments. We also present an end-to-end process for inferring a connected pedestrian path network map using street network information and our proposed dataset. The process features the use of a multi-input segmentation network trained on our dataset to predict important classes in the pedestrian environment and then generate a connected pedestrian path network. Our results demonstrate that the dataset is sufficiently large to train common segmentation models yielding accurate, robust pedestrian path networks.

show abstract

“…The up-to-date representative road extraction methods are based on deep learning, and the first attempt at deep learning based road extraction can be traced back to 2010 [10]. However, a decade of development has mainly focused on the development of specific convolutional neural network (CNN) structures for pixel-level road semantic segmentation [11][12][13][14][15][16][17][18][19][20][21][22], and how to regularize the extracted road surface maps to reach human-level and vector-based delineation has not been tackled in depth. In this paper, we propose a practical regularized road surface map extraction method based on a combined CNN and graph neural network (GNN) and the aid of road centerline maps, aiming to replace the use of human labor by directly predicting regularized and smooth double-line road vector maps.…”

Section: Introductionmentioning

confidence: 99%

“…Wang et al [21] proposed the nonlocal LinkNet model with non-local blocks in the encoder part to segment road surfaces from VHR satellite images. Zhou et al [22] proposed a network which embeds a universal iteration reinforcement module to enhance the learning ability.…”

Section: Introductionmentioning

confidence: 99%

A Combination of Convolutional and Graph Neural Networks for Regularized Road Surface Extraction

Yan

Yao

2022

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

Road surface extraction from high-resolution remote sensing images has many engineering applications; however, extracting regularized and smooth road surface maps that reach the human delineation level is a very challenging task, and substantial and time-consuming manual work is usually unavoidable. In this paper, to solve this problem, we propose a novel regularized road surface extraction framework by introducing a graph neural network (GNN) for processing the road graph that is preconstructed from the easily accessible road centerlines. The proposed framework formulates the road surface extraction problem as two-sided width inference of the road graph, and consists of a convolutional neural network (CNN)-based feature extractor and a GNN model for vertex attribute adjustment. The CNN extracts the high-level abstract features of each vertex in the graph as the input of the GNN, and also the road boundary features that allow us to distinguish roads from the background. The GNN propagates and aggregates the features of the vertices in the graph to achieve global optimization of the regression of the regularized widths of the vertices. At the same time, a biased centerline map can also be corrected based on the width prediction result. To the best of the authors' knowledge, this is the first study to have introduced a GNN to regularized human-level road surface extraction. The proposed method was evaluated on four diverse datasets, and the results show that the proposed method comprehensively outperforms the recent CNN-based segmentation methods and other regularization methods in the intersection over union (IoU) and smoothness score, and a visual check shows that a majority of the prediction results of the proposed method approach the human delineation level.

show abstract

FuNet: A Novel Road Extraction Network with Fusion of Location Data and Remote Sensing Imagery

Cited by 24 publications

References 47 publications

Road segmentation: exploiting the efficiency of skip connections for efficient semantic segmentation

Road segmentation: exploiting the efficiency of skip connections for efficient semantic segmentation

APE: An Open and Shared Annotated Dataset for Learning Urban Pedestrian Path Networks

A Combination of Convolutional and Graph Neural Networks for Regularized Road Surface Extraction

Contact Info

Product

Resources

About