Blind restoration of solar images via the Channel Sharing Spatio-temporal Network

Context. Due to the presence of atmospheric turbulence, the quality of solar images tends to be significantly degraded when observed by ground-based telescopes. The adaptive optics (AO) system can achieve partial correction but stops short of reaching the diffraction limit. In order to further impro...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Astronomy and astrophysics (Berlin) 2021-08, Vol.652, p.A50
Hauptverfasser: Wang, Shuai, Chen, Qingqing, He, Chunyuan, Zhang, Chi, Zhong, Libo, Bao, Hua, Rao, Changhui
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page A50
container_title Astronomy and astrophysics (Berlin)
container_volume 652
creator Wang, Shuai
Chen, Qingqing
He, Chunyuan
Zhang, Chi
Zhong, Libo
Bao, Hua
Rao, Changhui
description Context. Due to the presence of atmospheric turbulence, the quality of solar images tends to be significantly degraded when observed by ground-based telescopes. The adaptive optics (AO) system can achieve partial correction but stops short of reaching the diffraction limit. In order to further improve the imaging quality, post-processing for AO closed-loop images is still necessary. Methods based on deep learning (DL) have been proposed for AO image reconstruction, but the most of them are based on the assumption that the point spread function is spatially invariant. Aims. Our goal is to construct clear solar images by using a sophisticated spatially variant end-to-end blind restoration network. Methods. The proposed channel sharing spatio-temporal network (CSSTN) consists of three sub-networks: a feature extraction network, channel sharing spatio-temporal filter adaptive network (CSSTFAN), and a reconstruction network (RN). First, CSSTFAN generates two filters adaptively according to features generated from three inputs. Then these filters are delivered to the proposed channel sharing filter adaptive convolutional layer in CSSTFAN to convolve with the previous or current step features. Finally, the convolved features are concatenated as input of RN to restore a clear image. Ultimately, CSSTN and the other three supervised DL methods are trained on the binding real 705 nm photospheric and 656 nm chromospheric AO correction images as well as the corresponding speckle reconstructed images. Results. The results of CSSTN, the three DL methods, and one classic blind deconvolution method evaluated on four test sets are shown. The imaging condition of the first photospheric and second chromospheric set is the same as training set, except for the different time given in the same hour. The imaging condition of the third chromospheric and fourth photospheric set is the same as the first and second, except for the Sun region and time. Our method restores clearer images and performs best in both the peak signal-to-noise ratio and contrast among these methods.
doi_str_mv 10.1051/0004-6361/202140376
format Article
fullrecord <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_journals_2567980054</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2567980054</sourcerecordid><originalsourceid>FETCH-LOGICAL-c322t-1e9f9b9e98bc4960744ebefd3fb1858a0ba39e0f272d12ae7b1ac0867b2daa903</originalsourceid><addsrcrecordid>eNo9kN9LwzAQx4MoOKd_gS8Bn-suP5o2jzp0CkOR6XNI2mTr7JqadIr_vSmTPR0Hn7v73gehawK3BHIyAwCeCSbIjAIlHFghTtCEcEYzKLg4RZMjcY4uYtymlpKSTdDbfdt0NQ42Dj7oofEd9g5H3-qAm51e24i_G42HjcXzje462-LVRoemW-NVP_LZYHd9Gm3xix1-fPi8RGdOt9Fe_dcp-nh8eJ8_ZcvXxfP8bplVjNIhI1Y6aaSVpam4FCkmt8a6mjlDyrzUYDSTFhwtaE2otoUhuoJSFIbWWktgU3Rz2NsH_7VP-dXW70OXTiqai0KWADlPFDtQVfAxButUH9Jf4VcRUKM7NZpRoxl1dMf-ALCrYag</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2567980054</pqid></control><display><type>article</type><title>Blind restoration of solar images via the Channel Sharing Spatio-temporal Network</title><source>Bacon EDP Sciences France Licence nationale-ISTEX-PS-Journals-PFISTEX</source><source>EDP Sciences</source><source>Elektronische Zeitschriftenbibliothek - Frei zugängliche E-Journals</source><creator>Wang, Shuai ; Chen, Qingqing ; He, Chunyuan ; Zhang, Chi ; Zhong, Libo ; Bao, Hua ; Rao, Changhui</creator><creatorcontrib>Wang, Shuai ; Chen, Qingqing ; He, Chunyuan ; Zhang, Chi ; Zhong, Libo ; Bao, Hua ; Rao, Changhui</creatorcontrib><description>Context. Due to the presence of atmospheric turbulence, the quality of solar images tends to be significantly degraded when observed by ground-based telescopes. The adaptive optics (AO) system can achieve partial correction but stops short of reaching the diffraction limit. In order to further improve the imaging quality, post-processing for AO closed-loop images is still necessary. Methods based on deep learning (DL) have been proposed for AO image reconstruction, but the most of them are based on the assumption that the point spread function is spatially invariant. Aims. Our goal is to construct clear solar images by using a sophisticated spatially variant end-to-end blind restoration network. Methods. The proposed channel sharing spatio-temporal network (CSSTN) consists of three sub-networks: a feature extraction network, channel sharing spatio-temporal filter adaptive network (CSSTFAN), and a reconstruction network (RN). First, CSSTFAN generates two filters adaptively according to features generated from three inputs. Then these filters are delivered to the proposed channel sharing filter adaptive convolutional layer in CSSTFAN to convolve with the previous or current step features. Finally, the convolved features are concatenated as input of RN to restore a clear image. Ultimately, CSSTN and the other three supervised DL methods are trained on the binding real 705 nm photospheric and 656 nm chromospheric AO correction images as well as the corresponding speckle reconstructed images. Results. The results of CSSTN, the three DL methods, and one classic blind deconvolution method evaluated on four test sets are shown. The imaging condition of the first photospheric and second chromospheric set is the same as training set, except for the different time given in the same hour. The imaging condition of the third chromospheric and fourth photospheric set is the same as the first and second, except for the Sun region and time. Our method restores clearer images and performs best in both the peak signal-to-noise ratio and contrast among these methods.</description><identifier>ISSN: 0004-6361</identifier><identifier>EISSN: 1432-0746</identifier><identifier>DOI: 10.1051/0004-6361/202140376</identifier><language>eng</language><publisher>Heidelberg: EDP Sciences</publisher><subject>Adaptive optics ; Atmospheric turbulence ; Feature extraction ; Ground-based observation ; Image contrast ; Image quality ; Image reconstruction ; Image restoration ; Machine learning ; Photosphere ; Point spread functions ; Post-processing ; Signal to noise ratio ; Space telescopes</subject><ispartof>Astronomy and astrophysics (Berlin), 2021-08, Vol.652, p.A50</ispartof><rights>Copyright EDP Sciences Aug 2021</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c322t-1e9f9b9e98bc4960744ebefd3fb1858a0ba39e0f272d12ae7b1ac0867b2daa903</citedby><cites>FETCH-LOGICAL-c322t-1e9f9b9e98bc4960744ebefd3fb1858a0ba39e0f272d12ae7b1ac0867b2daa903</cites><orcidid>0000-0002-7897-2024 ; 0000-0001-8571-8502</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>314,776,780,3714,27901,27902</link.rule.ids></links><search><creatorcontrib>Wang, Shuai</creatorcontrib><creatorcontrib>Chen, Qingqing</creatorcontrib><creatorcontrib>He, Chunyuan</creatorcontrib><creatorcontrib>Zhang, Chi</creatorcontrib><creatorcontrib>Zhong, Libo</creatorcontrib><creatorcontrib>Bao, Hua</creatorcontrib><creatorcontrib>Rao, Changhui</creatorcontrib><title>Blind restoration of solar images via the Channel Sharing Spatio-temporal Network</title><title>Astronomy and astrophysics (Berlin)</title><description>Context. Due to the presence of atmospheric turbulence, the quality of solar images tends to be significantly degraded when observed by ground-based telescopes. The adaptive optics (AO) system can achieve partial correction but stops short of reaching the diffraction limit. In order to further improve the imaging quality, post-processing for AO closed-loop images is still necessary. Methods based on deep learning (DL) have been proposed for AO image reconstruction, but the most of them are based on the assumption that the point spread function is spatially invariant. Aims. Our goal is to construct clear solar images by using a sophisticated spatially variant end-to-end blind restoration network. Methods. The proposed channel sharing spatio-temporal network (CSSTN) consists of three sub-networks: a feature extraction network, channel sharing spatio-temporal filter adaptive network (CSSTFAN), and a reconstruction network (RN). First, CSSTFAN generates two filters adaptively according to features generated from three inputs. Then these filters are delivered to the proposed channel sharing filter adaptive convolutional layer in CSSTFAN to convolve with the previous or current step features. Finally, the convolved features are concatenated as input of RN to restore a clear image. Ultimately, CSSTN and the other three supervised DL methods are trained on the binding real 705 nm photospheric and 656 nm chromospheric AO correction images as well as the corresponding speckle reconstructed images. Results. The results of CSSTN, the three DL methods, and one classic blind deconvolution method evaluated on four test sets are shown. The imaging condition of the first photospheric and second chromospheric set is the same as training set, except for the different time given in the same hour. The imaging condition of the third chromospheric and fourth photospheric set is the same as the first and second, except for the Sun region and time. Our method restores clearer images and performs best in both the peak signal-to-noise ratio and contrast among these methods.</description><subject>Adaptive optics</subject><subject>Atmospheric turbulence</subject><subject>Feature extraction</subject><subject>Ground-based observation</subject><subject>Image contrast</subject><subject>Image quality</subject><subject>Image reconstruction</subject><subject>Image restoration</subject><subject>Machine learning</subject><subject>Photosphere</subject><subject>Point spread functions</subject><subject>Post-processing</subject><subject>Signal to noise ratio</subject><subject>Space telescopes</subject><issn>0004-6361</issn><issn>1432-0746</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2021</creationdate><recordtype>article</recordtype><recordid>eNo9kN9LwzAQx4MoOKd_gS8Bn-suP5o2jzp0CkOR6XNI2mTr7JqadIr_vSmTPR0Hn7v73gehawK3BHIyAwCeCSbIjAIlHFghTtCEcEYzKLg4RZMjcY4uYtymlpKSTdDbfdt0NQ42Dj7oofEd9g5H3-qAm51e24i_G42HjcXzje462-LVRoemW-NVP_LZYHd9Gm3xix1-fPi8RGdOt9Fe_dcp-nh8eJ8_ZcvXxfP8bplVjNIhI1Y6aaSVpam4FCkmt8a6mjlDyrzUYDSTFhwtaE2otoUhuoJSFIbWWktgU3Rz2NsH_7VP-dXW70OXTiqai0KWADlPFDtQVfAxButUH9Jf4VcRUKM7NZpRoxl1dMf-ALCrYag</recordid><startdate>20210801</startdate><enddate>20210801</enddate><creator>Wang, Shuai</creator><creator>Chen, Qingqing</creator><creator>He, Chunyuan</creator><creator>Zhang, Chi</creator><creator>Zhong, Libo</creator><creator>Bao, Hua</creator><creator>Rao, Changhui</creator><general>EDP Sciences</general><scope>AAYXX</scope><scope>CITATION</scope><scope>8FD</scope><scope>H8D</scope><scope>L7M</scope><orcidid>https://orcid.org/0000-0002-7897-2024</orcidid><orcidid>https://orcid.org/0000-0001-8571-8502</orcidid></search><sort><creationdate>20210801</creationdate><title>Blind restoration of solar images via the Channel Sharing Spatio-temporal Network</title><author>Wang, Shuai ; Chen, Qingqing ; He, Chunyuan ; Zhang, Chi ; Zhong, Libo ; Bao, Hua ; Rao, Changhui</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c322t-1e9f9b9e98bc4960744ebefd3fb1858a0ba39e0f272d12ae7b1ac0867b2daa903</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2021</creationdate><topic>Adaptive optics</topic><topic>Atmospheric turbulence</topic><topic>Feature extraction</topic><topic>Ground-based observation</topic><topic>Image contrast</topic><topic>Image quality</topic><topic>Image reconstruction</topic><topic>Image restoration</topic><topic>Machine learning</topic><topic>Photosphere</topic><topic>Point spread functions</topic><topic>Post-processing</topic><topic>Signal to noise ratio</topic><topic>Space telescopes</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Wang, Shuai</creatorcontrib><creatorcontrib>Chen, Qingqing</creatorcontrib><creatorcontrib>He, Chunyuan</creatorcontrib><creatorcontrib>Zhang, Chi</creatorcontrib><creatorcontrib>Zhong, Libo</creatorcontrib><creatorcontrib>Bao, Hua</creatorcontrib><creatorcontrib>Rao, Changhui</creatorcontrib><collection>CrossRef</collection><collection>Technology Research Database</collection><collection>Aerospace Database</collection><collection>Advanced Technologies Database with Aerospace</collection><jtitle>Astronomy and astrophysics (Berlin)</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Wang, Shuai</au><au>Chen, Qingqing</au><au>He, Chunyuan</au><au>Zhang, Chi</au><au>Zhong, Libo</au><au>Bao, Hua</au><au>Rao, Changhui</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Blind restoration of solar images via the Channel Sharing Spatio-temporal Network</atitle><jtitle>Astronomy and astrophysics (Berlin)</jtitle><date>2021-08-01</date><risdate>2021</risdate><volume>652</volume><spage>A50</spage><pages>A50-</pages><issn>0004-6361</issn><eissn>1432-0746</eissn><abstract>Context. Due to the presence of atmospheric turbulence, the quality of solar images tends to be significantly degraded when observed by ground-based telescopes. The adaptive optics (AO) system can achieve partial correction but stops short of reaching the diffraction limit. In order to further improve the imaging quality, post-processing for AO closed-loop images is still necessary. Methods based on deep learning (DL) have been proposed for AO image reconstruction, but the most of them are based on the assumption that the point spread function is spatially invariant. Aims. Our goal is to construct clear solar images by using a sophisticated spatially variant end-to-end blind restoration network. Methods. The proposed channel sharing spatio-temporal network (CSSTN) consists of three sub-networks: a feature extraction network, channel sharing spatio-temporal filter adaptive network (CSSTFAN), and a reconstruction network (RN). First, CSSTFAN generates two filters adaptively according to features generated from three inputs. Then these filters are delivered to the proposed channel sharing filter adaptive convolutional layer in CSSTFAN to convolve with the previous or current step features. Finally, the convolved features are concatenated as input of RN to restore a clear image. Ultimately, CSSTN and the other three supervised DL methods are trained on the binding real 705 nm photospheric and 656 nm chromospheric AO correction images as well as the corresponding speckle reconstructed images. Results. The results of CSSTN, the three DL methods, and one classic blind deconvolution method evaluated on four test sets are shown. The imaging condition of the first photospheric and second chromospheric set is the same as training set, except for the different time given in the same hour. The imaging condition of the third chromospheric and fourth photospheric set is the same as the first and second, except for the Sun region and time. Our method restores clearer images and performs best in both the peak signal-to-noise ratio and contrast among these methods.</abstract><cop>Heidelberg</cop><pub>EDP Sciences</pub><doi>10.1051/0004-6361/202140376</doi><orcidid>https://orcid.org/0000-0002-7897-2024</orcidid><orcidid>https://orcid.org/0000-0001-8571-8502</orcidid><oa>free_for_read</oa></addata></record>
fulltext fulltext
identifier ISSN: 0004-6361
ispartof Astronomy and astrophysics (Berlin), 2021-08, Vol.652, p.A50
issn 0004-6361
1432-0746
language eng
recordid cdi_proquest_journals_2567980054
source Bacon EDP Sciences France Licence nationale-ISTEX-PS-Journals-PFISTEX; EDP Sciences; Elektronische Zeitschriftenbibliothek - Frei zugängliche E-Journals
subjects Adaptive optics
Atmospheric turbulence
Feature extraction
Ground-based observation
Image contrast
Image quality
Image reconstruction
Image restoration
Machine learning
Photosphere
Point spread functions
Post-processing
Signal to noise ratio
Space telescopes
title Blind restoration of solar images via the Channel Sharing Spatio-temporal Network
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-10T07%3A31%3A38IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Blind%20restoration%20of%20solar%20images%20via%20the%20Channel%20Sharing%20Spatio-temporal%20Network&rft.jtitle=Astronomy%20and%20astrophysics%20(Berlin)&rft.au=Wang,%20Shuai&rft.date=2021-08-01&rft.volume=652&rft.spage=A50&rft.pages=A50-&rft.issn=0004-6361&rft.eissn=1432-0746&rft_id=info:doi/10.1051/0004-6361/202140376&rft_dat=%3Cproquest_cross%3E2567980054%3C/proquest_cross%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2567980054&rft_id=info:pmid/&rfr_iscdi=true