Loading [a11y]/accessibility-menu.js
Change Detection in Heterogeneous Satellite Remote-Sensing Imagery Using Deep Features Comparison | IEEE Conference Publication | IEEE Xplore

Change Detection in Heterogeneous Satellite Remote-Sensing Imagery Using Deep Features Comparison


Abstract:

This paper presents a novel method for detecting changes in heterogeneous multi-temporal remote sensing (RS) images using deep features comparison while exploiting two de...Show More

Abstract:

This paper presents a novel method for detecting changes in heterogeneous multi-temporal remote sensing (RS) images using deep features comparison while exploiting two deep learning models: Generative Adversarial Networks (GANs) and autoencoders. First, Deep Convolutional GANs are used to perform a pixel-to-pixel image translation to obtain two images of the same modality. Autoencoders are then trained and used to learn a compressed representation of the two images. Finally, a change map is obtained by combining/fusing the original image with its corresponding generated change-free image resulting from the difference between the two learned compressed representations. The method is general, able to accommodate change detection (CD) algorithms for RS images expressing any type of change. Quantitative and qualitative analysis of the proposed method applied to CD in very high resolution Synthetic Aperture Radar (SAR) and optical imagery show added value of generative DL models applied to CD in heterogeneous RS images as compared to existing CD methods.
Date of Conference: 07-12 July 2024
Date Added to IEEE Xplore: 05 September 2024
ISBN Information:

ISSN Information:

Conference Location: Athens, Greece

References

References is not available for this document.