DEADALUS: Massive-scale urban reconstuction, classification, and rendering from remote sensor imagery

遥感技术的进步使人们能够广泛获得地理图像，包括卫星和倾斜航空图像，以及一种称为广域运动图像的相对较新但迅速成熟的技术。遥感已经成功地应用于情报、安全、侦察、城市规划和监测等领域。然而，与城市区域对应的图像通常会产生大量数据，这就需要新的计算机视觉和三维图形技术来开发新兴的数据，包括三维重建、地理空间特征分类和重建模型的照片级真实感绘制。我们提出的研究方法是开发新的技术来处理包含重复模式的大数据量图像，使用密集匹配和深度学习进行3D重建、对象分类和逼真的外观建模。DRDC和Presagis都看到了康科迪亚研究人员进行的初步调查，Presagis是一家加拿大公司，在为航空航天、国防和汽车行业提供模拟、建模和VR软件方面处于世界领先地位，三家合作伙伴都认为坚持和遵循这一研究方法非常有益。DRDC和Presagis拥有广泛的领域知识，而Concordia的教职员工拥有必要的研究背景和专业知识。此外，Presagis拥有强大的软件工具，可以帮助创建训练深层网络所需的地面真实数据，并验证3D重建结果。该项目将促进对三名博士生和四名硕士生进行前沿学科的培训，如超大型遥感图像数据处理、3D图形、计算机视觉和深度学习，所有这些都是加拿大科学界和工业界非常感兴趣和受益的。基础研究和调查将为视觉、图形学和空中模拟的研究开辟新的领域。这些结果对加拿大来说是有益的，因为预计Presagis将能够利用它们向其世界各地的客户提供更快、更好和具有全球竞争力的解决方案。

Advances in remote sensing technologies have enabled the widespread availability of geo-imagery including satellite and oblique aerial images, and a relatively recent yet fast maturing technology called wide-area motion imagery (WAMI). Remote sensing has already been successfully employed in many applications such as intelligence, security, reconnaissance, urban planning and monitoring, etc. However, images corresponding to urban areas typically result in large volumes of data which require new computer vision and 3D graphics techniques for emerging data exploitation, including 3D reconstruction, geospatial feature classification and photorealistic rendering of the reconstructed models. Our proposed research methodology is to develop new techniques for handling big sets of large-sized images containing repetitive patterns using dense matching and deep learning for 3D reconstruction, object classification, and realistic appearance modeling. Preliminary investigations carried out by Concordia researchers have been seen by both DRDC and Presagis, a Canadian company and world leader in providing simulation, modeling, and VR software to aerospace, defense, and automotive industries, and all three partners consider it highly beneficial to persist and pursue with this research methodology. DRDC and Presagis have extensive domain knowledge, and Concordia faculty have the requisite research background and expertise. Further, Presagis has powerful software tools which will help create ground truth data needed for training deep networks, and also in validation of 3D reconstruction results. The project will facilitate training of three Doctoral and four Masters students in leading edge topics such as very large remote sensor image data processing, 3D graphics, computer vision, and deep learning, all of great interest and benefit to the Canadian scientific and industrial communities. The fundamental research andinvestigations should open new ground for research in vision, graphics and airborne simulation. The results are of benefit to Canada, as it is expected that Presagis, would be able to use them to provide faster, better and globally competitive solutions to its clients worldwide.