FusionRF: High-Fidelity Satellite Neural Radiance Fields from Multispectral and Panchromatic Acquisitions

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Sprintson, Michael, Chellappa, Rama, Peng, Cheng
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866909887208882176
author Sprintson, Michael
Chellappa, Rama
Peng, Cheng
author_facet Sprintson, Michael
Chellappa, Rama
Peng, Cheng
contents We introduce FusionRF, a novel framework for digital surface reconstruction from satellite multispectral and panchromatic images. Current work has demonstrated the increased accuracy of neural photogrammetry for surface reconstruction from optical satellite images compared to algorithmic methods. Common satellites produce both a panchromatic and multispectral image, which contain high spatial and spectral information respectively. Current neural reconstruction methods require multispectral images to be upsampled with a pansharpening method using the spatial data in the panchromatic image. However, these methods may introduce biases and hallucinations due to domain gaps. FusionRF introduces joint image fusion during optimization through a novel cross-resolution kernel that learns to resolve spatial resolution loss present in multispectral images. As input, FusionRF accepts the original multispectral and panchromatic data, eliminating the need for image preprocessing. FusionRF also leverages multimodal appearance embeddings that encode the image characteristics of each modality and view within a uniform representation. By optimizing on both modalities, FusionRF learns to fuse image modalities while performing reconstruction tasks and eliminates the need for a pansharpening preprocessing step. We evaluate our method on multispectral and panchromatic satellite images from the WorldView-3 satellite in various locations, and show that FusionRF provides an average of 17% reduction in depth reconstruction error, and renders sharp training and novel views.
format Preprint
id arxiv_https___arxiv_org_abs_2409_15132
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle FusionRF: High-Fidelity Satellite Neural Radiance Fields from Multispectral and Panchromatic Acquisitions
Sprintson, Michael
Chellappa, Rama
Peng, Cheng
Computer Vision and Pattern Recognition
Image and Video Processing
We introduce FusionRF, a novel framework for digital surface reconstruction from satellite multispectral and panchromatic images. Current work has demonstrated the increased accuracy of neural photogrammetry for surface reconstruction from optical satellite images compared to algorithmic methods. Common satellites produce both a panchromatic and multispectral image, which contain high spatial and spectral information respectively. Current neural reconstruction methods require multispectral images to be upsampled with a pansharpening method using the spatial data in the panchromatic image. However, these methods may introduce biases and hallucinations due to domain gaps. FusionRF introduces joint image fusion during optimization through a novel cross-resolution kernel that learns to resolve spatial resolution loss present in multispectral images. As input, FusionRF accepts the original multispectral and panchromatic data, eliminating the need for image preprocessing. FusionRF also leverages multimodal appearance embeddings that encode the image characteristics of each modality and view within a uniform representation. By optimizing on both modalities, FusionRF learns to fuse image modalities while performing reconstruction tasks and eliminates the need for a pansharpening preprocessing step. We evaluate our method on multispectral and panchromatic satellite images from the WorldView-3 satellite in various locations, and show that FusionRF provides an average of 17% reduction in depth reconstruction error, and renders sharp training and novel views.
title FusionRF: High-Fidelity Satellite Neural Radiance Fields from Multispectral and Panchromatic Acquisitions
topic Computer Vision and Pattern Recognition
Image and Video Processing
url https://arxiv.org/abs/2409.15132