Semantic segmentation of roads in high-resolution satellite imagery

Mina Al Saad, Nour Aburaed, Saeed Al Mansoori, Hussain Al Ahmad, Stephen Marshall

Research output: Contribution to conferencePaperpeer-review

Abstract

Remote sensing is a field that is constantly growing due to its importance in applications related to urban planning, risk assessment, resource exploration, and disaster management. Some of the most useful image processing tasks in the area of remote sensing include object detection, scene classification, and semantic segmentation. Nowadays, researchers are looking into automating these tasks, as performing them manually is time and cost inefficient. Therefore, Deep Learning (DL) algorithms have been utilized for remote sensing and image processing tasks in the recent years. Many studies proved efficiency of DL algorithms, however, several challenges remain unsolved. In particular, semantic segmentation is one of the challenging tasks that is under heavy demand in the field of remote sensing. For this research study, semantic segmentation is applied to extract roads from high-resolution satellite images. Road mapping is an essential first step in several applications that include transportation, traffic management, and city planning. Extracting roads efficiently and accurately will in turn boost the overall outcome of the aforementioned applications. U-Net is a state-of-the-art Convolutional Neural Networks (CNN) that will be used to perform semantic segmentation. U-Net will be trained using a publicly available dataset, which is Massachusetts Road. Afterwards, their performance is enhanced by introducing extra feature layers and finding proper balance between down-sampling trade off and accurate object boundary localization, which is a current open problem in the area of semantic segmentation. Samples results will be demonstrated and evaluated in terms of the overall accuracy, loss, and dice coefficient.

Original languageEnglish
Publication statusPublished - 14 Oct 2020
Event71st International Astronautical Congress - Virtual
Duration: 12 Oct 202014 Oct 2020
https://www.iafastro.org/events/iac/iac-2020/

Conference

Conference71st International Astronautical Congress
Abbreviated titleIAC 2020
Period12/10/2014/10/20
Internet address

Keywords

  • convolutional neural network
  • road extraction
  • satellite images
  • semantic segmentation

Fingerprint

Dive into the research topics of 'Semantic segmentation of roads in high-resolution satellite imagery'. Together they form a unique fingerprint.

Cite this