Constructing 3D Models of Rigid Objects from Satellite Images with High Spatial Resolution Using Convolutional Neural Networks
— A way of constructing 3D models of rigid objects from one satellite image is described. It is based on the use of two convolution neural networks which sequentially process high-resolution satellite images. The first neural network performs integral image analysis for segmentation and identificati...
Gespeichert in:
| Veröffentlicht in: | Izvestiya. Atmospheric and oceanic physics 2020-12, Vol.56 (12), p.1664-1677 |
|---|---|
| Hauptverfasser: | , , , , |
| Format: | Artikel |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | 1677 |
|---|---|
| container_issue | 12 |
| container_start_page | 1664 |
| container_title | Izvestiya. Atmospheric and oceanic physics |
| container_volume | 56 |
| creator | Gvozdev, O. G. Kozub, V. A. Kosheleva, N. V. Murynin, A. B. Richter, A. A. |
| description | —
A way of constructing 3D models of rigid objects from one satellite image is described. It is based on the use of two convolution neural networks which sequentially process high-resolution satellite images. The first neural network performs integral image analysis for segmentation and identification of objects of specified physical classes. The second neural network performs local image analysis and works with images segmented by the first neural network in areas of the image that presumably contain objects of specified classes. An algorithm for reconstructing a 3D model of an object from raster domains of a segmented image obtained from local analysis is described. It is based on regression analysis, the assessing of equivalent figures, and the linearization and polarization of contours. Results from the algorithm’s operation are given using the example of railway infrastructure facilities. The results from constructing 3D models of three objects of the railway infrastructure, identified via the operation of neural networks for four informative classes of areas are presented, e.g. roofs, walls, railroad tracks, contanc lines (poles). Standard dimensions (e.g., the railway gauge (1.52 m) and the height of railway support poles (11.35 m)) are used to estimate scaling coefficients that allow determination of base dimensions and object heights. The possibility of constructing 3D models of objects of areas from 210 to 4200 m
2
is shown. |
| doi_str_mv | 10.1134/S0001433820120427 |
| format | Article |
| fullrecord | <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_journals_2499403983</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2499403983</sourcerecordid><originalsourceid>FETCH-LOGICAL-c316t-55383943d9c38c5d5babfd94d1f03cc843600f4f3354487fe81d3ccdf4e648313</originalsourceid><addsrcrecordid>eNp1UE1LwzAYDqLgnP4AbwHP1aRv0qVHmR8bTAebA2-lS5Mus2tmkjq8-Ntt3cCDeHrg-XpfHoQuKbmmFNjNnBBCGYCICY0JiwdHqEc551ESi9dj1OvkqNNP0Zn3a0KSmJFBD30Nbe2Da2QwdYnhDj_ZQlUeW41npjQFni7XSgaPtbMbPM-DqioTFB5v8lJ5vDNhhUemXOH5Ng8mr_BMeVs1wdgaL3zX2R74ODCt_Kwa9wNhZ92bP0cnOq-8ujhgHy0e7l-Go2gyfRwPbyeRBJqEiHMQkDIoUglC8oIv86UuUlZQTUBKwSAhRDMNwBkTA60ELVq-0EwlTACFPrra926dfW-UD9naNq59yGcxS1NGIBXQuujeJZ313imdbZ3Z5O4zoyTrZs7-zNxm4n3Gt966VO63-f_QN6G6f5g</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2499403983</pqid></control><display><type>article</type><title>Constructing 3D Models of Rigid Objects from Satellite Images with High Spatial Resolution Using Convolutional Neural Networks</title><source>SpringerLink Journals - AutoHoldings</source><creator>Gvozdev, O. G. ; Kozub, V. A. ; Kosheleva, N. V. ; Murynin, A. B. ; Richter, A. A.</creator><creatorcontrib>Gvozdev, O. G. ; Kozub, V. A. ; Kosheleva, N. V. ; Murynin, A. B. ; Richter, A. A.</creatorcontrib><description>—
A way of constructing 3D models of rigid objects from one satellite image is described. It is based on the use of two convolution neural networks which sequentially process high-resolution satellite images. The first neural network performs integral image analysis for segmentation and identification of objects of specified physical classes. The second neural network performs local image analysis and works with images segmented by the first neural network in areas of the image that presumably contain objects of specified classes. An algorithm for reconstructing a 3D model of an object from raster domains of a segmented image obtained from local analysis is described. It is based on regression analysis, the assessing of equivalent figures, and the linearization and polarization of contours. Results from the algorithm’s operation are given using the example of railway infrastructure facilities. The results from constructing 3D models of three objects of the railway infrastructure, identified via the operation of neural networks for four informative classes of areas are presented, e.g. roofs, walls, railroad tracks, contanc lines (poles). Standard dimensions (e.g., the railway gauge (1.52 m) and the height of railway support poles (11.35 m)) are used to estimate scaling coefficients that allow determination of base dimensions and object heights. The possibility of constructing 3D models of objects of areas from 210 to 4200 m
2
is shown.</description><identifier>ISSN: 0001-4338</identifier><identifier>EISSN: 1555-628X</identifier><identifier>DOI: 10.1134/S0001433820120427</identifier><language>eng</language><publisher>Moscow: Pleiades Publishing</publisher><subject>Algorithms ; Artificial neural networks ; Climatology ; Coefficients ; Convolution ; Dimensions ; Earth and Environmental Science ; Earth Sciences ; Geophysics/Geodesy ; High resolution ; Image analysis ; Image processing ; Image reconstruction ; Image resolution ; Image segmentation ; Infrastructure ; Neural networks ; Poles ; Railway construction ; Railway tracks ; Regression analysis ; Resolution ; Satellite imagery ; Satellites ; Scaling ; Spaceborne remote sensing ; Spatial discrimination ; Spatial resolution ; Three dimensional models ; Ways and Means of Processing and Interpreting Space-Based Information</subject><ispartof>Izvestiya. Atmospheric and oceanic physics, 2020-12, Vol.56 (12), p.1664-1677</ispartof><rights>Pleiades Publishing, Ltd. 2020. ISSN 0001-4338, Izvestiya, Atmospheric and Oceanic Physics, 2020, Vol. 56, No. 12, pp. 1664–1677. © Pleiades Publishing, Ltd., 2020. Russian Text © The Author(s), 2020, published in Issledovanie Zemli iz Kosmosa, 2020, No. 5, pp. 78–96.</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c316t-55383943d9c38c5d5babfd94d1f03cc843600f4f3354487fe81d3ccdf4e648313</citedby><cites>FETCH-LOGICAL-c316t-55383943d9c38c5d5babfd94d1f03cc843600f4f3354487fe81d3ccdf4e648313</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://link.springer.com/content/pdf/10.1134/S0001433820120427$$EPDF$$P50$$Gspringer$$H</linktopdf><linktohtml>$$Uhttps://link.springer.com/10.1134/S0001433820120427$$EHTML$$P50$$Gspringer$$H</linktohtml><link.rule.ids>314,780,784,27922,27923,41486,42555,51317</link.rule.ids></links><search><creatorcontrib>Gvozdev, O. G.</creatorcontrib><creatorcontrib>Kozub, V. A.</creatorcontrib><creatorcontrib>Kosheleva, N. V.</creatorcontrib><creatorcontrib>Murynin, A. B.</creatorcontrib><creatorcontrib>Richter, A. A.</creatorcontrib><title>Constructing 3D Models of Rigid Objects from Satellite Images with High Spatial Resolution Using Convolutional Neural Networks</title><title>Izvestiya. Atmospheric and oceanic physics</title><addtitle>Izv. Atmos. Ocean. Phys</addtitle><description>—
A way of constructing 3D models of rigid objects from one satellite image is described. It is based on the use of two convolution neural networks which sequentially process high-resolution satellite images. The first neural network performs integral image analysis for segmentation and identification of objects of specified physical classes. The second neural network performs local image analysis and works with images segmented by the first neural network in areas of the image that presumably contain objects of specified classes. An algorithm for reconstructing a 3D model of an object from raster domains of a segmented image obtained from local analysis is described. It is based on regression analysis, the assessing of equivalent figures, and the linearization and polarization of contours. Results from the algorithm’s operation are given using the example of railway infrastructure facilities. The results from constructing 3D models of three objects of the railway infrastructure, identified via the operation of neural networks for four informative classes of areas are presented, e.g. roofs, walls, railroad tracks, contanc lines (poles). Standard dimensions (e.g., the railway gauge (1.52 m) and the height of railway support poles (11.35 m)) are used to estimate scaling coefficients that allow determination of base dimensions and object heights. The possibility of constructing 3D models of objects of areas from 210 to 4200 m
2
is shown.</description><subject>Algorithms</subject><subject>Artificial neural networks</subject><subject>Climatology</subject><subject>Coefficients</subject><subject>Convolution</subject><subject>Dimensions</subject><subject>Earth and Environmental Science</subject><subject>Earth Sciences</subject><subject>Geophysics/Geodesy</subject><subject>High resolution</subject><subject>Image analysis</subject><subject>Image processing</subject><subject>Image reconstruction</subject><subject>Image resolution</subject><subject>Image segmentation</subject><subject>Infrastructure</subject><subject>Neural networks</subject><subject>Poles</subject><subject>Railway construction</subject><subject>Railway tracks</subject><subject>Regression analysis</subject><subject>Resolution</subject><subject>Satellite imagery</subject><subject>Satellites</subject><subject>Scaling</subject><subject>Spaceborne remote sensing</subject><subject>Spatial discrimination</subject><subject>Spatial resolution</subject><subject>Three dimensional models</subject><subject>Ways and Means of Processing and Interpreting Space-Based Information</subject><issn>0001-4338</issn><issn>1555-628X</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2020</creationdate><recordtype>article</recordtype><recordid>eNp1UE1LwzAYDqLgnP4AbwHP1aRv0qVHmR8bTAebA2-lS5Mus2tmkjq8-Ntt3cCDeHrg-XpfHoQuKbmmFNjNnBBCGYCICY0JiwdHqEc551ESi9dj1OvkqNNP0Zn3a0KSmJFBD30Nbe2Da2QwdYnhDj_ZQlUeW41npjQFni7XSgaPtbMbPM-DqioTFB5v8lJ5vDNhhUemXOH5Ng8mr_BMeVs1wdgaL3zX2R74ODCt_Kwa9wNhZ92bP0cnOq-8ujhgHy0e7l-Go2gyfRwPbyeRBJqEiHMQkDIoUglC8oIv86UuUlZQTUBKwSAhRDMNwBkTA60ELVq-0EwlTACFPrra926dfW-UD9naNq59yGcxS1NGIBXQuujeJZ313imdbZ3Z5O4zoyTrZs7-zNxm4n3Gt966VO63-f_QN6G6f5g</recordid><startdate>20201201</startdate><enddate>20201201</enddate><creator>Gvozdev, O. G.</creator><creator>Kozub, V. A.</creator><creator>Kosheleva, N. V.</creator><creator>Murynin, A. B.</creator><creator>Richter, A. A.</creator><general>Pleiades Publishing</general><general>Springer Nature B.V</general><scope>AAYXX</scope><scope>CITATION</scope><scope>7TG</scope><scope>7TN</scope><scope>F1W</scope><scope>H96</scope><scope>KL.</scope><scope>L.G</scope></search><sort><creationdate>20201201</creationdate><title>Constructing 3D Models of Rigid Objects from Satellite Images with High Spatial Resolution Using Convolutional Neural Networks</title><author>Gvozdev, O. G. ; Kozub, V. A. ; Kosheleva, N. V. ; Murynin, A. B. ; Richter, A. A.</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c316t-55383943d9c38c5d5babfd94d1f03cc843600f4f3354487fe81d3ccdf4e648313</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2020</creationdate><topic>Algorithms</topic><topic>Artificial neural networks</topic><topic>Climatology</topic><topic>Coefficients</topic><topic>Convolution</topic><topic>Dimensions</topic><topic>Earth and Environmental Science</topic><topic>Earth Sciences</topic><topic>Geophysics/Geodesy</topic><topic>High resolution</topic><topic>Image analysis</topic><topic>Image processing</topic><topic>Image reconstruction</topic><topic>Image resolution</topic><topic>Image segmentation</topic><topic>Infrastructure</topic><topic>Neural networks</topic><topic>Poles</topic><topic>Railway construction</topic><topic>Railway tracks</topic><topic>Regression analysis</topic><topic>Resolution</topic><topic>Satellite imagery</topic><topic>Satellites</topic><topic>Scaling</topic><topic>Spaceborne remote sensing</topic><topic>Spatial discrimination</topic><topic>Spatial resolution</topic><topic>Three dimensional models</topic><topic>Ways and Means of Processing and Interpreting Space-Based Information</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Gvozdev, O. G.</creatorcontrib><creatorcontrib>Kozub, V. A.</creatorcontrib><creatorcontrib>Kosheleva, N. V.</creatorcontrib><creatorcontrib>Murynin, A. B.</creatorcontrib><creatorcontrib>Richter, A. A.</creatorcontrib><collection>CrossRef</collection><collection>Meteorological & Geoastrophysical Abstracts</collection><collection>Oceanic Abstracts</collection><collection>ASFA: Aquatic Sciences and Fisheries Abstracts</collection><collection>Aquatic Science & Fisheries Abstracts (ASFA) 2: Ocean Technology, Policy & Non-Living Resources</collection><collection>Meteorological & Geoastrophysical Abstracts - Academic</collection><collection>Aquatic Science & Fisheries Abstracts (ASFA) Professional</collection><jtitle>Izvestiya. Atmospheric and oceanic physics</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Gvozdev, O. G.</au><au>Kozub, V. A.</au><au>Kosheleva, N. V.</au><au>Murynin, A. B.</au><au>Richter, A. A.</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Constructing 3D Models of Rigid Objects from Satellite Images with High Spatial Resolution Using Convolutional Neural Networks</atitle><jtitle>Izvestiya. Atmospheric and oceanic physics</jtitle><stitle>Izv. Atmos. Ocean. Phys</stitle><date>2020-12-01</date><risdate>2020</risdate><volume>56</volume><issue>12</issue><spage>1664</spage><epage>1677</epage><pages>1664-1677</pages><issn>0001-4338</issn><eissn>1555-628X</eissn><abstract>—
A way of constructing 3D models of rigid objects from one satellite image is described. It is based on the use of two convolution neural networks which sequentially process high-resolution satellite images. The first neural network performs integral image analysis for segmentation and identification of objects of specified physical classes. The second neural network performs local image analysis and works with images segmented by the first neural network in areas of the image that presumably contain objects of specified classes. An algorithm for reconstructing a 3D model of an object from raster domains of a segmented image obtained from local analysis is described. It is based on regression analysis, the assessing of equivalent figures, and the linearization and polarization of contours. Results from the algorithm’s operation are given using the example of railway infrastructure facilities. The results from constructing 3D models of three objects of the railway infrastructure, identified via the operation of neural networks for four informative classes of areas are presented, e.g. roofs, walls, railroad tracks, contanc lines (poles). Standard dimensions (e.g., the railway gauge (1.52 m) and the height of railway support poles (11.35 m)) are used to estimate scaling coefficients that allow determination of base dimensions and object heights. The possibility of constructing 3D models of objects of areas from 210 to 4200 m
2
is shown.</abstract><cop>Moscow</cop><pub>Pleiades Publishing</pub><doi>10.1134/S0001433820120427</doi><tpages>14</tpages></addata></record> |
| fulltext | fulltext |
| identifier | ISSN: 0001-4338 |
| ispartof | Izvestiya. Atmospheric and oceanic physics, 2020-12, Vol.56 (12), p.1664-1677 |
| issn | 0001-4338 1555-628X |
| language | eng |
| recordid | cdi_proquest_journals_2499403983 |
| source | SpringerLink Journals - AutoHoldings |
| subjects | Algorithms Artificial neural networks Climatology Coefficients Convolution Dimensions Earth and Environmental Science Earth Sciences Geophysics/Geodesy High resolution Image analysis Image processing Image reconstruction Image resolution Image segmentation Infrastructure Neural networks Poles Railway construction Railway tracks Regression analysis Resolution Satellite imagery Satellites Scaling Spaceborne remote sensing Spatial discrimination Spatial resolution Three dimensional models Ways and Means of Processing and Interpreting Space-Based Information |
| title | Constructing 3D Models of Rigid Objects from Satellite Images with High Spatial Resolution Using Convolutional Neural Networks |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-10T04%3A02%3A11IST&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=Constructing%203D%20Models%20of%20Rigid%20Objects%20from%20Satellite%20Images%20with%20High%20Spatial%20Resolution%20Using%20Convolutional%20Neural%20Networks&rft.jtitle=Izvestiya.%20Atmospheric%20and%20oceanic%20physics&rft.au=Gvozdev,%20O.%20G.&rft.date=2020-12-01&rft.volume=56&rft.issue=12&rft.spage=1664&rft.epage=1677&rft.pages=1664-1677&rft.issn=0001-4338&rft.eissn=1555-628X&rft_id=info:doi/10.1134/S0001433820120427&rft_dat=%3Cproquest_cross%3E2499403983%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=2499403983&rft_id=info:pmid/&rfr_iscdi=true |