Semantic Segmentation and Edge Detection—Approach to Road Detection in Very High Resolution Satellite Images
Road detection technology plays an essential role in a variety of applications, such as urban planning, map updating, traffic monitoring and automatic vehicle navigation. Recently, there has been much development in detecting roads in high-resolution (HR) satellite images based on semantic segmentat...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-01-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/14/3/613 |
| _version_ | 1797485014848372736 |
|---|---|
| author | Hamza Ghandorh Wadii Boulila Sharjeel Masood Anis Koubaa Fawad Ahmed Jawad Ahmad |
| author_facet | Hamza Ghandorh Wadii Boulila Sharjeel Masood Anis Koubaa Fawad Ahmed Jawad Ahmad |
| author_sort | Hamza Ghandorh |
| collection | DOAJ |
| description | Road detection technology plays an essential role in a variety of applications, such as urban planning, map updating, traffic monitoring and automatic vehicle navigation. Recently, there has been much development in detecting roads in high-resolution (HR) satellite images based on semantic segmentation. However, the objects being segmented in such images are of small size, and not all the information in the images is equally important when making a decision. This paper proposes a novel approach to road detection based on semantic segmentation and edge detection. Our approach aims to combine these two techniques to improve road detection, and it produces sharp-pixel segmentation maps, using the segmented masks to generate road edges. In addition, some well-known architectures, such as SegNet, used multi-scale features without refinement; thus, using attention blocks in the encoder to predict fine segmentation masks resulted in finer edges. A combination of weighted cross-entropy loss and the focal Tversky loss as the loss function is also used to deal with the highly imbalanced dataset. We conducted various experiments on two datasets describing real-world datasets covering the three largest regions in Saudi Arabia and Massachusetts. The results demonstrated that the proposed method of encoding HR feature maps effectively predicts sharp segmentation masks to facilitate accurate edge detection, even against a harsh and complicated background. |
| first_indexed | 2024-03-09T23:13:43Z |
| format | Article |
| id | doaj.art-792a1cf857c5400183b115853c92598e |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-09T23:13:43Z |
| publishDate | 2022-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-792a1cf857c5400183b115853c92598e2023-11-23T17:40:38ZengMDPI AGRemote Sensing2072-42922022-01-0114361310.3390/rs14030613Semantic Segmentation and Edge Detection—Approach to Road Detection in Very High Resolution Satellite ImagesHamza Ghandorh0Wadii Boulila1Sharjeel Masood2Anis Koubaa3Fawad Ahmed4Jawad Ahmad5College of Computer Science and Engineering, Taibah University, Medina 42353, Saudi ArabiaRobotics and Internet-of-Things Laboratory, Prince Sultan University, Riyadh 12435, Saudi ArabiaHealthHub, Seoul 06524, KoreaRobotics and Internet-of-Things Laboratory, Prince Sultan University, Riyadh 12435, Saudi ArabiaDepartment of Cyber Security, Pakistan Navy Engineering College, NUST, Islamabad 75350, PakistanSchool of Computing, Edinburgh Napier University, Edinburgh EH10 5DT, UKRoad detection technology plays an essential role in a variety of applications, such as urban planning, map updating, traffic monitoring and automatic vehicle navigation. Recently, there has been much development in detecting roads in high-resolution (HR) satellite images based on semantic segmentation. However, the objects being segmented in such images are of small size, and not all the information in the images is equally important when making a decision. This paper proposes a novel approach to road detection based on semantic segmentation and edge detection. Our approach aims to combine these two techniques to improve road detection, and it produces sharp-pixel segmentation maps, using the segmented masks to generate road edges. In addition, some well-known architectures, such as SegNet, used multi-scale features without refinement; thus, using attention blocks in the encoder to predict fine segmentation masks resulted in finer edges. A combination of weighted cross-entropy loss and the focal Tversky loss as the loss function is also used to deal with the highly imbalanced dataset. We conducted various experiments on two datasets describing real-world datasets covering the three largest regions in Saudi Arabia and Massachusetts. The results demonstrated that the proposed method of encoding HR feature maps effectively predicts sharp segmentation masks to facilitate accurate edge detection, even against a harsh and complicated background.https://www.mdpi.com/2072-4292/14/3/613deep learningconvolutional neural networks2D attentionsatellite imagesroad segmentationedge detection |
| spellingShingle | Hamza Ghandorh Wadii Boulila Sharjeel Masood Anis Koubaa Fawad Ahmed Jawad Ahmad Semantic Segmentation and Edge Detection—Approach to Road Detection in Very High Resolution Satellite Images Remote Sensing deep learning convolutional neural networks 2D attention satellite images road segmentation edge detection |
| title | Semantic Segmentation and Edge Detection—Approach to Road Detection in Very High Resolution Satellite Images |
| title_full | Semantic Segmentation and Edge Detection—Approach to Road Detection in Very High Resolution Satellite Images |
| title_fullStr | Semantic Segmentation and Edge Detection—Approach to Road Detection in Very High Resolution Satellite Images |
| title_full_unstemmed | Semantic Segmentation and Edge Detection—Approach to Road Detection in Very High Resolution Satellite Images |
| title_short | Semantic Segmentation and Edge Detection—Approach to Road Detection in Very High Resolution Satellite Images |
| title_sort | semantic segmentation and edge detection approach to road detection in very high resolution satellite images |
| topic | deep learning convolutional neural networks 2D attention satellite images road segmentation edge detection |
| url | https://www.mdpi.com/2072-4292/14/3/613 |
| work_keys_str_mv | AT hamzaghandorh semanticsegmentationandedgedetectionapproachtoroaddetectioninveryhighresolutionsatelliteimages AT wadiiboulila semanticsegmentationandedgedetectionapproachtoroaddetectioninveryhighresolutionsatelliteimages AT sharjeelmasood semanticsegmentationandedgedetectionapproachtoroaddetectioninveryhighresolutionsatelliteimages AT aniskoubaa semanticsegmentationandedgedetectionapproachtoroaddetectioninveryhighresolutionsatelliteimages AT fawadahmed semanticsegmentationandedgedetectionapproachtoroaddetectioninveryhighresolutionsatelliteimages AT jawadahmad semanticsegmentationandedgedetectionapproachtoroaddetectioninveryhighresolutionsatelliteimages |