Abstract
Availability of very high-resolution remote sensing images and advancement of deep learning methods have shifted the paradigm of image classification from pixel-based and object-based methods to deep learning-based semantic segmentation. This shift demands a structured analysis and revision of the current status on the research domain of deep learning-based semantic segmentation. The focus of this paper is on urban remote sensing images. We review and perform a meta-analysis to juxtapose recent papers in terms of research problems, data source, data preparation methods including pre-processing and augmentation techniques, training details on architectures, backbones, frameworks, optimizers, loss functions and other hyper-parameters and performance comparison. Our detailed review and meta-analysis show that deep learning not only outperforms traditional methods in terms of accuracy, but also addresses several challenges previously faced. Further, we provide future directions of research in this domain.
Highlights
IntroductionLand-use/land-cover (LULC) maps are often generated from medium resolution satellite images like Sentinel [1] and Landsat [2]
We present detailed and tabulated analysis and discuss how the recent papers have tried to address the challenges faced during the shift of paradigm to semantic segmentation such as: ignorance of spatial/contextual information by convolutional neural networks (CNN), boundary pixel classification problems, class imbalance problem, domain-shift problem, salt-and-pepper noise, structural stereotype, insufficient training and other limitations
We will discuss the improvements achieved with the help of traditional methods (Section 4.1), improvements shown by Deep learning (DL) (Section 4.2) and dataset challenges (Section 4.3) and research problems addressed by DL (Section 4.4)
Summary
Land-use/land-cover (LULC) maps are often generated from medium resolution satellite images like Sentinel [1] and Landsat [2]. These images are useful to classify land cover classes like built area, residential area, vegetation surface, impervious surface, water, etc. A finer-resolution image pixel is more useful than a greater number of spectral bands or narrower interval of wavelength [3]. This is the reason why commercial satellite images and unmanned aerial vehicles (UAV) are more popular as they aim to increase the visibility of terrestrial objects, especially urban features, by reducing per-pixel size. With the increase in spatial resolution, more urban objects are clearly visible in satellite images, and studies shifted its paradigm from spectral image classification, pixel-based image analysis (PBIA) and object-based image analysis (OBIA)
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
