Road Extraction Methods in High-Resolution Remote Sensing Images: A Comprehensive Review

Automatic road extraction from very-high-resolution remote sensing images has become a popular topic in a wide range of fields. Convolutional neural networks are often used for this purpose. However, many network models do not achieve satisfactory extraction results because of the elongated nature and varying sizes of roads in images. To improve the accuracy of road extraction, this paper proposes a deep learning model based on the structure of Deeplab v3. It incorporates squeeze-and-excitation (SE) module to apply weights to different feature channels, and performs multi-scale upsampling to preserve and fuse shallow and deep information. To solve the problems associated with unbalanced road samples in images, different loss functions and backbone network modules are tested in the model’s training process. Compared with cross entropy, dice loss can improve the performance of the model during training and prediction. The SE module is superior to ResNext and ResNet in improving the integrity of the extracted roads. Experimental results obtained using the Massachusetts Roads Dataset show that the proposed model (Nested SE-Deeplab) improves F1-Score by 2.4% and Intersection over Union by 2.0% compared with FC-DenseNet. The proposed model also achieves better segmentation accuracy in road extraction compared with other mainstream deep-learning models including Deeplab v3, SegNet, and UNet.

Download Full-text

Building Extraction from High-Resolution Remote Sensing Images Based on GrabCut with Automatic Selection of Foreground and Background Samples

Photogrammetric Engineering & Remote Sensing ◽

10.14358/pers.86.4.235 ◽

2020 ◽

Vol 86 (4) ◽

pp. 235-245 ◽

Cited By ~ 1

Author(s):

Ka Zhang ◽

Hui Chen ◽

Wen Xiao ◽

Yehua Sheng ◽

Dong Su ◽

...

Keyword(s):

Remote Sensing ◽

High Resolution ◽

Geodesic Distance ◽

Extraction Methods ◽

Building Extraction ◽

Remote Sensing Images ◽

Canny Operator ◽

Contour Lines ◽

Average Accuracy ◽

Seed Points

This article proposes a new building extraction method from high-resolution remote sensing images, based on GrabCut, which can automatically select foreground and background samples under the constraints of building elevation contour lines. First the image is rotated according to the direction of pixel displacement calculated by the rational function Model. Second, the Canny operator, combined with morphology and the Hough transform, is used to extract the building's elevation contour lines. Third, seed points and interesting points of the building are selected under the constraint of the contour line and the geodesic distance. Then foreground and background samples are obtained according to these points. Fourth, GrabCut and geometric features are used to carry out image segmentation and extract buildings. Finally, WorldView satellite images are used to verify the proposed method. Experimental results show that the average accuracy can reach 86.34%, which is 15.12% higher than other building extraction methods.

Download Full-text

A Fast Algorithm for Road Recognition in Remote Sensing Image

Advanced Materials Research ◽

10.4028/www.scientific.net/amr.108-111.1344 ◽

2010 ◽

Vol 108-111 ◽

pp. 1344-1347

Author(s):

Li Li Li ◽

Yong Xin Liu

Keyword(s):

Remote Sensing ◽

Edge Detection ◽

Fast Algorithm ◽

Remote Sensing Image ◽

Extraction Methods ◽

Feature Integration ◽

Road Extraction ◽

Remote Sensing Images ◽

The Road ◽

Simulation Results

In general, the road extraction methods in remote sensing images mainly are edge detection, feature integration, and so on. A fast road recognition arithmetic is presented in this paper. First using adaptive binarization arithmetic, the path on remote sensing images is extracted. Then morphological method is used to process image. Finally, the extracted image superimposed with the original and get clear road. Simulation results shows that this algorithm is efficiency, the anti-noise ability is enhance, and more precision.

Download Full-text

Road extraction from high-resolution remote sensing images based on characteristics

10.1117/12.749423 ◽

2007 ◽

Author(s):

Jie Yu ◽

Huiling Qin ◽

Qin Yan ◽

Ming Tan ◽

Guoning Zhang

Keyword(s):

Remote Sensing ◽

High Resolution ◽

Road Extraction ◽

Remote Sensing Images

Download Full-text

Road extraction from high-resolution remote sensing images using wavelet transform and hough transform

2012 5th International Congress on Image and Signal Processing ◽

10.1109/cisp.2012.6469890 ◽

2012 ◽

Cited By ~ 3

Author(s):

Xiaoliang Yang ◽

Gongjian Wen

Keyword(s):

Remote Sensing ◽

Wavelet Transform ◽

High Resolution ◽

Hough Transform ◽

Road Extraction ◽

Remote Sensing Images

Download Full-text

Road Extraction of High-Resolution Remote Sensing Images Derived from DenseUNet

Remote Sensing ◽

10.3390/rs11212499 ◽

2019 ◽

Vol 11 (21) ◽

pp. 2499 ◽

Cited By ~ 6

Author(s):

Jiang Xin ◽

Xinchang Zhang ◽

Zhiqiang Zhang ◽

Wu Fang

Keyword(s):

Remote Sensing ◽

High Resolution ◽

Intelligent Transportation Systems ◽

Road Network ◽

Semantic Segmentation ◽

Transportation Systems ◽

Machine Learning Algorithms ◽

Road Extraction ◽

Remote Sensing Images ◽

Road Network Extraction

Road network extraction is one of the significant assignments for disaster emergency response, intelligent transportation systems, and real-time updating road network. Road extraction base on high-resolution remote sensing images has become a hot topic. Presently, most of the researches are based on traditional machine learning algorithms, which are complex and computational because of impervious surfaces such as roads and buildings that are discernible in the images. Given the above problems, we propose a new method to extract the road network from remote sensing images using a DenseUNet model with few parameters and robust characteristics. DenseUNet consists of dense connection units and skips connections, which strengthens the fusion of different scales by connections at various network layers. The performance of the advanced method is validated on two datasets of high-resolution images by comparison with three classical semantic segmentation methods. The experimental results show that the method can be used for road extraction in complex scenes.

Download Full-text

Based on High Resolution of Remote Sensing Data Mining Houses Information Extraction Methods Research

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.170-173.2803 ◽

2012 ◽

Vol 170-173 ◽

pp. 2803-2807

Author(s):

Yan Hua Sun ◽

Ping Wang

Keyword(s):

Remote Sensing ◽

High Resolution ◽

Information Extraction ◽

Spatial Resolution ◽

Mining Area ◽

Extraction Methods ◽

Accuracy Evaluation ◽

Remote Sensing Images ◽

Extraction Technology ◽

Solar Altitude

High resolution remote sensing images generally refer to image to the spatial resolution within 10m aerospace、aviation remote sensing images. The emergence of high-resolution images strengthened the ability to recognize the large scale features, especially for the extraction of houses information in mining area. High spatial resolution image has rich delicate texture feature, it is urgent to solution the problem of how to extract the features. The technology is very useful for statistic houses information、village relocation assessment and research of pressure coal status, providing important data basis for village relocation, statistics, assessment. Taking henan as a mining area for example, houses information extraction methods are discussed. This paper mainly research contents as followings: It is combined with the space texture information of high resolution imaging rich, using different methods to extract building information, including followings: First, ordinary image segmentation technology; this method is simple and feasible, but extracted housing information is not accurate. Second, the object-oriented method of feature extraction technology, visualization degree and extracting accuracy of this method is higher; Third, it has conducted the preliminary height extraction of the houses; according to the solar altitude angles and the shadow of the houses to calculate the height of the houses. And considering the influence of undulating terrain, using the terrain DEM data to analyze study area, finally determined the shadow length, and then used solar altitude angles to calculate houses height. Based on the verification, accuracy evaluation results show that houses contour information extraction accuracy is: accuracy of the number and area is over 80%, the total rate of wrong classifications is lower. Houses highly information extraction accuracy is within the 85%. The research methods are effective.

Download Full-text