Abstract
A new convolution neural network (CNN) architecture for semantic segmentation of high resolution aerial imagery is proposed in this paper. The proposed architecture follows an hourglass-shaped network (HSN) design being structured into encoding and decoding stages. By taking advantage of recent advances in CNN designs, we use the composed inception module to replace common convolutional layers, providing the network with multi-scale receptive areas with rich context. Additionally, in order to reduce spatial ambiguities in the up-sampling stage, skip connections with residual units are also employed to feed forward encoding-stage information directly to the decoder. Moreover, overlap inference is employed to alleviate boundary effects occurring when high resolution images are inferred from small-sized patches. Finally, we also propose a post-processing method based on weighted belief propagation to visually enhance the classification results. Extensive experiments based on the Vaihingen and Potsdam datasets demonstrate that the proposed architectures outperform three reference state-of-the-art network designs both numerically and visually.
Highlights
Semantic segmentation in remote sensing aims at accurately labeling each pixel in an aerial image by assigning it to a specific class, such as vegetation, buildings, vehicles or roads
We focus on the semantic segmentation of high-resolution aerial images and propose a convolution neural network (CNN)-based solution by following this generic design methodology for supervised-learning
For FPL [22], we have carried out experiments using the original FPL network, which was trained on the Vainhingen and Potsdam datasets and was publicly made available by its authors
Summary
Semantic segmentation in remote sensing aims at accurately labeling each pixel in an aerial image by assigning it to a specific class, such as vegetation, buildings, vehicles or roads. This is a very important task that facilitates a wide set of applications ranging from urban planning to change detection and automated-map making [1]. One of the major challenges is given by the ever-increasing spatial and spectral resolution of remote sensing images. High spectral resolutions provide abundant information for Earth observations, but selecting, fusing and classifying hyperspectral images remain significant research challenges in remote sensing [2,3]
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.
