Abstract
Regularized Building Boundary Extraction From Remote Sensing Imagery Based on Augment Feature Pyramid Network and Morphological Constraint
Highlights
Building in urban structures is one of the most critical components of ground information
The method of patched convolutional neural networks (CNN) can improve the accuracy of building extraction, it cannot take into account the information inside and outside the patch, which leads to the limited ability of building extraction and cannot maximize the role of deep learning method [25]
1) We propose an augment feature pyramid network for building extraction, which can provide more accurate and intensive global features for semantic segmentation tasks and effectively reduce the loss of information
Summary
Building in urban structures is one of the most critical components of ground information. Drăgu et al used eCognition software for object-oriented classification of remote sensing images based on the characteristics of the building, including image context, geometric structure, and spectral features, which can effectively reduce noise and improve accuracy [7]. This type of method can achieve high accuracy, it cannot automatically determine the optimal segmentation scale parameter and is too subjective. Wu et al combined with multi-scale features to fulfill boundary extraction based on an improved U-shaped network [30] They analyzed the characteristics of Batch Normalization (BN), Relu, and LeakyRelu in detail.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
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 callMore From: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Disclaimer: 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.
