Hervé Lahamy

Structural Deflection Measurement with a Range Camera

AbstractRange cameras offer great potential for the measurement of structural deformations because of their ability to directly measure video sequences of three-dimensional coordinates of entire surfaces, their compactness, and their relatively low cost compared with other active imaging technologies such as terrestrial laser scanners. Identified limitations of range cameras for high-precision metrology applications such as deformation measurement include the high (centimeter level) noise level and scene-dependent errors. This paper proposes models and methodologies to overcome these limitations and reports on the use of a SwissRanger SR4000 range camera for the measurement of deflections in concrete beams subjected to flexural load-testing. Results from three separate tests show that submillimeter precision and accuracy—assessed by comparison with estimates derived from terrestrial laser scanner data—can be achieved. The high-accuracy range camera results were realized by eliminating the systematic, scen...

Towards Real-Time and Rotation-Invariant American Sign Language Alphabet Recognition Using a Range Camera

The automatic interpretation of human gestures can be used for a natural interaction with computers while getting rid of mechanical devices such as keyboards and mice. In order to achieve this objective, the recognition of hand postures has been studied for many years. However, most of the literature in this area has considered 2D images which cannot provide a full description of the hand gestures. In addition, a rotation-invariant identification remains an unsolved problem, even with the use of 2D images. The objective of the current study was to design a rotation-invariant recognition process while using a 3D signature for classifying hand postures. A heuristic and voxel-based signature has been designed and implemented. The tracking of the hand motion is achieved with the Kalman filter. A unique training image per posture is used in the supervised classification. The designed recognition process, the tracking procedure and the segmentation algorithm have been successfully evaluated. This study has demonstrated the efficiency of the proposed rotation invariant 3D hand posture signature which leads to 93.88% recognition rate after testing 14,732 samples of 12 postures taken from the alphabet of the American Sign Language.

Haptic and gesture-based interactions for manipulating geological datasets

In this paper, we report the first step of our ongoing research toward the creation of an intuitive and interactive environment for manipulating and analyzing geological datasets. This first step of our project aimed at the development of a manipulation system through the employment of haptic sense and gesture detection, into a virtual environment. The developed prototype integrates stereoscopic interactive visual rendering, haptic feedback and real-time hand tracking via a range camera, in a multithreaded architecture. Our prototype has been tested with both a synthetic 3D terrain and geological datasets. Our first results confirm the effectiveness of the selected architecture.

Measurement of Deflection in Concrete Beams During Fatigue Loading Test Using the Microsoft Kinect 2.0

Abstract This study focuses on 3 Hz fatigue load testing of a reinforced concrete beam in laboratory conditions. Three-dimensional (3D) image time series of the beam’s top surface were captured with the Microsoft time-of-flight Kinect 2.0 sensor. To estimate the beam deflection, the imagery was first segmented to extract the top surface of the beam. The centre line was then modeled using third-order B-splines. The deflection of the beam as a function of time was estimated from the modeled centre line and, following past practice, also at several witness plates attached to the side of the beam. Subsequent correlation of the peak displacement with the applied loading cycles permitted estimation of fatigue in the beam. The accuracy of the deflections was evaluated by comparison with the measurements obtained using a Keyence LK-G407 laser displacement sensors. The results indicate that the deflections can be recovered with sub-millimetre accuracy using the centreline profile modelling method.

Evaluating the capability of time-of-flight cameras for accurately imaging a cyclically loaded beam

Time-of-flight cameras are used for diverse applications ranging from human-machine interfaces and gaming to robotics and earth topography. This paper aims at evaluating the capability of the Mesa Imaging SR4000 and the Microsoft Kinect 2.0 time-of-flight cameras for accurately imaging the top surface of a concrete beam subjected to fatigue loading in laboratory conditions. Whereas previous work has demonstrated the success of such sensors for measuring the response at point locations, the aim here is to measure the entire beam surface in support of the overall objective of evaluating the effectiveness of concrete beam reinforcement with steel fibre reinforced polymer sheets. After applying corrections for lens distortions to the data and differencing images over time to remove systematic errors due to internal scattering, the periodic deflections experienced by the beam have been estimated for the entire top surface of the beam and at witness plates attached. The results have been assessed by comparison with measurements from highly-accurate laser displacement transducers. This study concludes that both the Microsoft Kinect 2.0 and the Mesa Imaging SR4000s are capable of sensing a moving surface with sub-millimeter accuracy once the image distortions have been modeled and removed.

Heuristic and voxel-based signature for hand posture recognition using a range camera

To improve the interaction between humans and machines, hand gestures have been a studied alternative for many years. Most of the literature in this area has considered 2D images which cannot provide a full description of the hand gestures due mainly to self occlusion. The objective of the current study is to increase the number of gestures recognizable in real-time while using a 3D signature. An heuristic and voxel-based signature has been designed and implemented. To evaluate the latter, an exhaustive performance analysis including comparison with ground truth and with other well-known features and classifiers was conducted. This study has demonstrated the efficiency of the proposed 3D hand posture signature which leads to 84% recognition rate after testing around 30000 samples of 18 gestures in real-time.

An Egg Volume Measurement System Based on the Microsoft Kinect

Measuring the volume of bird eggs is a very important task for the poultry industry and ornithological research due to the high revenue generated by the industry. In this paper, we describe a prototype of a new metrological system comprising a 3D range camera, Microsoft Kinect (Version 2) and a point cloud post-processing algorithm for the estimation of the egg volume. The system calculates the egg volume directly from the egg shape parameters estimated from the least-squares method in which the point clouds of eggs captured by the Kinect are fitted to novel geometric models of an egg in a 3D space. Using the models, the shape parameters of an egg are estimated along with the egg’s position and orientation simultaneously under the least-squares criterion. Four sets of experiments were performed to verify the functionality and the performance of the system, while volumes estimated from the conventional water displacement method and the point cloud captured by a survey-grade laser scanner serve as references. The results suggest that the method is straightforward, feasible and reliable with an average egg volume estimation accuracy 93.3% when compared to the reference volumes. As a prototype, the software part of the system was implemented in a post-processing mode. However, as the proposed processing techniques is computationally efficient, the prototype can be readily transformed into a real-time egg volume system.

Performance analysis of different classification methods for hand gesture recognition using range cameras

Most of the methods described in the literature for automatic hand gesture recognition make use of classification techniques with a variety of features and classifiers. This research focuses on the frequently-used ones by performing a comparative analysis using datasets collected with a range camera. Eight different gestures were considered in this research. The features include Hu-moments, orientation histograms and hand shape associated with its distance transformation image. As classifiers, the k-nearest neighbor algorithm and the chamfer distance have been chosen. For an extensive comparison, four different databases have been collected with variation in translation, orientation and scale. The evaluation has been performed by measuring the separability of classes, and by analyzing the overall recognition rates as well as the processing times. The best result is obtained from the combination of the chamfer distance classifier and hand shape and distance transformation image, but the time analysis reveals that the corresponding processing time is not adequate for a real-time recognition.