Short CV

Arnold Irschara is a research assistant at the Institute for Computer Graphics and Vision, Graz University of Technology. He received a MSc degree in Telematics from Graz University of Technology in 2006.

Research Interests

• Structure from Motion (SfM)
• Large scale 3D reconstruction
• Image-based localization
• (UAVs) Unmanned Aerial Vehicles
• Aerial Photogrammetry
• Image retrieval and object recognition

Click here for a list of publications.

Projects

3D Reconstruction using UAVs

Efficient Structure from Motion with Weak Position and Orientation Priors

In this paper we present an approach that leverages prior information from global positioning systems and inertial measurement units to speedup structure from motion computation. We propose a view selection strategy that advances vocabulary tree based coarse matching by also considering the geometric configuration between weakly oriented images. Furthermore, we introduce a fast and scalable reconstruction approach that relies on global rotation registration and robust bundle adjustment. Real world experiments are performed using data acquired by a micro aerial vehicle attached with GPS/INS sensors. Our proposed algorithm achieves orientation results that are sub-pixel accurate and the precision is on a par with results from incremental structure from motion approaches. Moreover, the method is scalable and computationally more efficient than previous approaches.

A. Irschara, C. Hoppe, S. Kluckner, H. Bischof. Workshop of Aerial Video Processing (WAVP) in conjunction with CVPR 2011. (PDF), Video

Towards Fully Automatic Photogrammetric Rreconstruction Using Digital Images Taken From UAVs

We argue that the future of remote sensing will see a diversification of sensors and sensor platforms. We argue further that remote sensing will also benefit from recent advances in computing technology to employ new algorithms previously too complex to apply. In this paper we support this argument by three demonstrations. First, we show that an unmanned aerial vehicle (UAV) equipped with digital cameras can provide valuable visual information about the Earth’s surface rapidly and at low cost from nearly any viewpoint. Second, we demonstrate an end-to-end workflow to process a sizeable block of such imagery in a fully automated manner. Thirdly, we build this workflow on a novel computing system taking advantage of the invention of the Graphics Processing Unit (GPU) that is capable of performing complex algorithms in an acceptable elapsed time. The transition to diverse imaging sensors and platforms results in a requirement to deal with unordered sets of images, such as typically collected from a UAV, and to match and orientate these images automatically. Our approach is fully automated and capable of addressing large datasets in reasonable time and at low costs on a standard desktop PC. We compare our method to a semi-automatic orientation approach based on the PhotoModeler software and demonstrate superior performance in terms of automation, accuracy and processing time.

A. Irschara, V.Kaufmann, M.Klopschitz, H. Bischof and F. Leberl. ISPRS 2010. (PDF), Video

Aerial Photogrammetry

Point Clouds: Lidar versus 3D Vision

Novel automated photogrammetry technologies are based on four innovations. First is the cost-free increase of overlaps between images when sensing is by digital means rather than film. Second is a vastly improved radiometric perfor- mance. Third is the transition towards multi-view matching and geometry. Fourth is the invention of the Graphics Processing Unit (GPU) that makes it possible to employ previously prohibitively complex algorithms for image matching. These innovations have led to an improved automation success of the photogrammetric workflow. Photogrammetry is therefore capable of producing point clouds at sub-pixel accuracy, at very dense point intervals and in near real-time, thereby eroding the traditional unique selling proposition of LiDAR scanners.

F. Leberl, A. Irschara, T. Pock, P. Meixner, M. Gruber, S. Scholz, and A. Wiechert. Photogrammetric Engineering and Remote Sensing,Vol. 76, No. 10, pp 1123-1134., 2010. (PDF), Video

Image-based Localization

From Structure-from-Motion Point Clouds to Fast Location Recognition

Efficient view registration with respect to a given 3D reconstruction has many applications like inside-out tracking in indoor and outdoor environments, and geo-locating images from large photo collections. We present a fast location recognition technique based on structure from motion point clouds. Vocabulary tree-based indexing of features directly returns relevant fragments of 3D models instead of documents from the images database. Additionally, we propose a compressed 3D scene representation which improves recognition rates while simultaneously reducing the computation time and the memory consumption. The design of our method is based on algorithms that efficiently utilize modern graphics processing units to deliver real-time performance for view registration. We demonstrate the approach by matching hand-held outdoor videos to known 3D urban models, and by registering images from online photo collections to the corresponding landmarks.

A. Irschara, C. Zach, J.-M. Frahm, H. Bischof. CVPR 2009. Preprint (PDF), Video

Structure from Motion

What Can Missing Correspondences Tell Us About 3D Structure and Motion?

Practically all existing approaches to structure and motion computation use only positive image correspondences to verify the camera pose hypotheses. Incorrect epipolar geometries are solely detected by identifying outliers among the found correspondences. Ambigous patterns in the images are often incorrectly handled by these standard methods. In this work we propose two approaches to overcome such problems. First, we apply non-monotone reasoning on view triplets using a Bayesian formulation. In contrast to two-view epipolar geometry, image triplets allow the prediction of features in the third image. Absence of these features (i.e. missing correspondences) enables additional inference about the view triplet. Furthermore, we integrate these view triplet handling into an incremental procedure for structure and motion computation. Thus, our approach is able to refine the maintained 3D structure when additional image data is provided.

C. Zach, A. Irschara, H. Bischof. CVPR 2008. Paper (PDF)

Towards Wiki-based Dense City Modeling

This work reports on the advances and on the current status of a terrestrial city modeling approach, which uses images contributed by end-users as input. Hence, the Wiki principle well known from textual knowledge databases is transferred to the goal of incrementally building a virtual representation of the occupied habitat. In order to achieve this objective, many state-of-the-art computer vision methods must be applied and modified according to this task. We describe the utilized 3D vision methods and show initial results obtained from the current image database acquired by in-house participants.

A. Irschara, C. Zach, H. Bischof. VRML Workshop in conjunction with ICCV 2007. Paper (PDF) Video 1 Video 2