精化多視角影像密匹配及點雲產製

Refinement of Multi-view Dense Image Matching and Point Cloud Generation

10.6574/JPRS.202106_26(2).0002

劉宣萱(Hsuan-Hsuan Liu)；趙鍵哲(Jen-Jer Jaw)

影像密匹配 ； 多視角 ； 點雲 ； 精化 ； Dense Image Matching ； Multi-view ； Point Cloud ； Refinement

航測及遙測學刊

26卷2期（2021 / 06 / 01）

75 - 94

繁體中文

大量多視角影像於密匹配計算處理上較為複雜而繁瑣，且針對每一像對分別計算其初始視差值，耗損的時間成本亦相對增加；再者，具多組像對重疊條件之多視角影像，倘未善加調製其交會幾何，產製之場景點雲即便具有描述幾何多餘觀測特性，然而點位的不精確性以及較大的離散度亦無助於後續空間資訊之應用。對此，本研究提出一系列優化作業模式並區分為三大主軸：建立影像群聚關係、視差傳遞策略和點雲精化策略等。業經兩組實際資料驗證其功效及可行性，說明所研擬方法產製之點雲能有效描述場景幾何，且於兩測試區域之像對計算總量部分，減少約為44％及14％，而針對其時效提升部分則達80％以上。

Processing large number of multi-view images is complicated and tedious. Also, when matching multiple stereo pairs, it would take long in getting disparity values if each pair is to be processed independently. In addition, redundantly described scene models with low reliability trouble the exploitation of geospatial information. This paper proposes an effective matching strategy featuring in key view selection and clustering, disparity delivery and point cloud refinement to tackle the aforementioned shortcomings. The proposed approach has been tested by two practical data sets and it is proven that both the efficient image manipulation with the computational time reduction more than 80% and quality point cloud generation well depicting the scene geometry highlight the merit of this study.

1. Ahmadabadian, A.H.,Robson, S.,Boehm, J.,Shortis, M.(2013).Image selection in photogrammetric multi-view stereo methods for metric and complete 3D reconstruction.Proceedings of the SPIE Vol.8791, Videometrics, Range Imaging, and Applications XII; and Automated Visual Inspection,Munich, Germany:
2. Hirschmüller, H.(2005).Accurate and efficient stereo processing by semi-global matching and mutual information.Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition,San Diego, CA, USA:
3. Hirschmüller, H.(2008).Stereo processing by semiglobal matching and mutual information.IEEE Transactions on Pattern Analysis and Machine Intelligence,30(2),328-341.
4. Mahami, H.,Nasirzadeh, F.,Ahmadabadian, A.H.,Esmaeili, F.,Nahavandi, S.(2019).Imaging network design to improve the automated construction progress monitoring process.Construction Innovation,19(3),386-404.
5. Messinger, M.,Silman, M.(2016).Unmanned aerial vehicles for the assessment and monitoring of environmental contamination: an example from coal ash spills.Environmental Pollution,218,889-894.
6. Rothermel, M.(2016).Germany, Stuttgart,University of Stuttgart.
7. Rothermel, M.,Gong, K.,Fritsch, D.,Schindler, K.,Haala, N.(2020).Photometric multi-view mesh refinement for high-resolution satellite images.ISPRS Journal of Photogrammetry and Remote Sensing,166,52-62.
8. Rothermel, M.,Wenzel, K.,Fritsch, D.,Haala, N.(2012).SURE: Photogrammetric surface reconstruction from imagery.Proceedings of LC3D Workshop,Berlin:
9. Topal, C.,Akinlar, C.(2012).Edge Drawing: A combined real-time edge and segment detector.Journal of Visual Communication and Image Representation,23(6),862-872.
10. Wenzel, K.,Rothermel, M.,Fritsch, D.,Haala, N.(2013).Image acquisition and model selection for multi-view stereo.Proceedings of the ISPRS International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences,40(5W),251-258.
11. Zhang, K.,Lu, J.B.,Lafruit, G.(2009).Cross-based matching local stereo matching using orthogonal integral images.IEEE Transactions on Circuits and Systems for Video Technology,19(7),1073-1079.
12. 丁皓偉, H.W.(2014)。Taiwan, ROC，國立臺灣大學土木工程學系=National Taiwan University。
13. 李欣錡, H.C.(2016)。Taiwan, ROC，國立臺灣大學土木工程學系=National Taiwan University。