Alexander Reiterer

A 3D optical deformation measurement system supported by knowledge-based and learning techniques

Abstract High accuracy 3D representation and monitoring of objects is receiving increasing interest both in science and industrial applications. Up to now tasks like monitoring of building displacements or deformations were solved by means of artificial targets on the objects of interest, although mature optical 3D measurement and laser scanning techniques are available. Such systems can perform their measurements even without targeting. This paper presents a new optical 3D measurement system, based on the fusion between a geodetic image sensor and a laser scanner. The main goal of its development was the automation of the whole measurement process, including the tasks of point identification and measurement, deformation analysis, and interpretation. This was only possible by means of new methods and techniques originally developed in the area of Artificial Intelligence; both point detection and deformation analysis are supported by decision systems that use such techniques. The resulting complex multi-sensor system is able to measure and analyse the deformation of objects, as shown in experiments. In this article we focus on specific key components and novel techniques that have been developed, and briefly report on the current stage of the whole system.

Track detection in 3D laser scanning data of railway infrastructure

Novel safety systems are needed to meet the growing demand of railway operation. In this paper we introduce general techniques for the detection of tracks and their components in 3D laser scanning data. These techniques make use of feature based methods, such as support vector machines, as well as model based methods, such as template matching. The focus of this work are robust and precise detectors for infrastructure elements, such as rails, tracks, closure rails, and frogs. These parts can be used for both, track maintenance and train-borne localization. The approach is evaluated experimentally on 3D laser scanning data and compared with a reference system. Furthermore, the approach is generic such that it can be used for data of any suitable laser scanning system.

Vision-Based Terrestrial Surface Monitoring

The monitoring of geo-risk areas is getting more and more importance due to increasing damage caused by hazardous events such as rock slides, as a result of the environmental change. Terrestrial long-range sensing (up to several kilometres of distance between sensor and target region) is a valuable means for monitoring such sites using non-signalized targets in high resolution, which is necessary to detect regions, amount, direction and trends of motion early enough to take risk mitigation measures. The technology to realize such a sensing strategy combines various fields of research, such as sensor technology, surveying, computer vision and geological sciences. This chapter describes two vision-based sensing techniques suited for terrestrial surface monitoring (terrestrial laser scanning, and image-based tacheometers), and their sensing strategies, data processing and data exploitation issues. Examples for monitoring frameworks are given, and technical and engineering solutions are described. A set of applications from permafrost, glacier and snow cover monitoring, as well as rock fall site monitoring shows the relevance, technologic maturity and limits of existing approaches. Rock falls and other geo-hazards being the major fields of application for such systems, the chances of saving lives, protecting infrastructure and habitats and avoiding injury to field personnel are increased so that the better and more accurate event can be monitored. The research and technology described in this chapter will help the surveying, photogrammetry and computer vision community fighting global warming impacts.

Quality Control for Building Industry by Means of a New Optical 3D Measurement and Analysis System

In the past, high-precision online 3D measuring required artificial targets defining the points on the objects to be monitored. For many tasks such as monitoring of displacements of buildings, artificial targets are not desired. Today’s image-assisted measurement systems can perform their measurements even without targeting. Such systems use the texture on the surface of the object to find “interesting points” which can replace the artificial targets. However, well-trained “measurement experts” are required to operate such a measurement system. In order to make such systems easy to use even for non-experts, it can be extended by an appropriate decision system which supports the operator. At the Institute of Geodesy and Geophysics of Vienna University of Technology a new kind of image-based measurement system is being developed (research project “Multi-Sensor Deformation Measurement System Supported by Knowledge-Based and Cognitive Vision Techniques”). This system can be used for measuring, analysing and interpreting deformations for the task of quality control in civil engineering. The work-flow is based on new techniques originally developed in the field of Artificial Intelligence

A Distance-Based Method for the Evaluation of Interest point Detection Algorithms

Many approaches in computer vision are based on point detection algorithms. In the literature, a wide variety of such algorithms are available. Therefore, it is an important task to evaluate existing and newly developed point detection algorithms. Up to now, this process was done by means of several methods. In this paper, we recall current point detection evaluation techniques, and motivated by their insufficiency for certain applications, we present a new application-oriented method which is based on distances between sets of points.

Image-based 3D surface approximation of the bladder using structure-from-motion for enhanced cystoscopy based on phantom data

Abstract Bladder cancer is likely to recur after resection. For this reason, bladder cancer survivors often undergo follow-up cystoscopy for years after treatment to look for bladder cancer recurrence. 3D modeling of the bladder could provide more reliable cystoscopic documentation by giving an overall picture of the organ and tumor positions. However, 3D reconstruction of the urinary bladder based on endoscopic images is challenging. This is due to the small field of view of the endoscope, considerable image distortion, and occlusion by urea, blood or particles. In this paper, we will demonstrate a method for the conversion of uncalibrated, monocular, endoscopic videos of the bladder into a 3D model using structure-from-motion (SfM). First of all, frames are extracted from video sequences. Distortions are then corrected in a calibration procedure. Finally, the 3D reconstruction algorithm generates a sparse surface approximation of the bladder lining based on the corrected frames. This method was tested using an endoscopic video of a phantom that mimics the rich structure of the bladder. The reconstructed 3D model covered a large part of the object, with an average reprojection error of 1.15 pixels and a relative accuracy of 99.4%.

Knowledge-Based Geo-risk Assessment for an Intelligent Measurement System

Rockfalls and landslides are major types of natural hazards worldwide that kill or injure a large number of individuals and cause very high costs every year. Risk assessment of such dangerous events requires an accurate evaluation of the geology, hydrogeology, morphology and interrelated factors such as environmental conditions and human activities. It is of particular importance for engineers and geologists to assess slope stability and dynamics in order to take appropriate, effective and timely measures against such events. This paper presents a decision-tool for geo-risk assessment on the basis of a knowledge-based system. The integration of such a tool with novel measurement sensors into an advanced system for geo-risk monitoring, which performs data fusion on-line, is innovative. To enable such a system, a knowledge base capturing domain knowledge formally is developed, which to the best of our knowledge is unprecedented; the completed part for initial risk assessment works quite well, as extensive experiments with a number of human experts have shown.

WiKaF - A Knowledge-based Kalman-Filter for Pedestrian Positioning

The precise, reliable and preferably ubiquitous positioning of mobile users is a substantial precondition for the provision of location and situation based information by Location Based Services. Within these applications, determining the location of pedestrians in ‘passive environment’ represents a special challenge. In this paper a new knowledge-based approach for the improvement of position quality is presented.