Semantic Segmentation and Edge Detection—Approach to Road Detection in Very High Resolution Satellite Images

Abstract

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.

Novel intelligent architecture and approximate solution for future networks

“…The suggested techniques presented in this paper can be exploited and operated to be adopted to the problems developed in [48][49][50][51][52]. We believe that our approach is important since it provides optimal security due to its multilevel security that minimizes the level of leaked information in case of cybersecurity attacks compared to other approaches.…”

Real time read-frequency optimization for railway monitoring system

“…The proposed algorithms can be improved and adopted to the problems studied in [4,9,10,17,37]. Several applications of scheduling problems in real life can be exploited to enhance the proposed algorithms [3, 21, 24-26, 29, 36].…”

TAU: A framework for video-based traffic analytics leveraging artificial intelligence and unmanned aerial systems

“…Road detection from UAV was the target of many studies in literature (Ghandorh et al, 2022;Zhou et al, 2014;Lin and Saripalli, 2012;Kim, 2005;Pless and Jurgens, 2004). The aim was to localize the areas where the different types of vehicles circulate.…”

“…On the other hand, Since 2012 (Krizhevsky et al, 2012), an impressive advancement in computer vision applications has been noted. This advancement is due to the appearance of the area of deep learning algorithms that made image and video processing more efficient than ever before (Goodfellow et al, 2016;Liu et al, 2020;Benjdira et al, 2019;Benjdira et al, 2020;Alhichri et al, 2019Alhichri et al, , 2016Koubâa et al, 2020b;Boulila et al, 2021;Alkhelaiwi et al, 2021;Ben Atitallah et al, 2022;Benjdira et al, 2022;Koubâa et al, 2020a;Noor et al, 2020;Ben Jabra et al, 2021). Many deep learning algorithms have been incrementally optimized for the main tasks on deep learning such as object detection inside images (Liu et al, 2020), object tracking in videos (Ciaparrone et al, 2020), instance segmentation (Hafiz and Bhat, 2020) and semantic segmentation (Hao et al, 2020).…”