Please use this identifier to cite or link to this item:
|Title:||Use of Low-Cost Digital Consumer Camera for Stereo-Photogrammetry of Structures Undergoing Destructive Tests|
|Authors:||LATHUILIERE Anne-Claire; CAPERAN PHILIPPE|
|Citation:||Proceedings of the 3rd Workshop on Optical Measurement Techniques for Structures and Systems (OPTIMESS) p. 1-7|
|Publisher:||OPTIMESS Scientific Research Network supported by the Fund for Scientific Research, Flanders|
|Type:||Articles in periodicals and books|
|Abstract:||The European Laboratory for Structural Assessment is able to test large structures on its reaction wall and strong floor. Conventional sensors are used to collect displacement, elongation and strain on particular points of the structures, but they are punctual, or integrate the information along lines. Furthermore, a typical experiment in ELSA involves huge forces applied to the structure, an action in antinomy with a perfect control of the boundary conditions. Beyond the fact that it permits to have field of observations, photogrammetry allows quick and accurate diagnostic of the setting of experiments, and characterisation of their boundary conditions. Thus, photogrammetry has been introduced in the laboratory complementarily to the traditional techniques. In some case, the structure is driven to a brittle ruin, which could possibly lead to the loss of costly sensors or camera. The study presented here is a first step to evaluate photogrammetry based on low-cost digital consumer camera. On the long term, it could permit to increase the number of low-cost camera while keeping information at a fair level. The experimental setting was composed of a cylindrical shell (length 6.7 m, diameter 55 cm) of composite material anchored on the reaction wall and progressively bent by applying a torque at its free extremity. The zone under scrutiny was a joint linking two cylindrical subassemblies of the structure. A net of 13x9 targets (squares of side 22 mm) was disposed on the carbon fibre joint with a mesh step of approximately 35 mm. The cameras were two webcams Philips ToUcam Pro with CCD sensor of 640x480 pixels on 8 bits. The stereo set-up was calibrated with the method of Bouguet. The targets were identified through morphological image analysis. Their temporal positions on series of 57 images pairs- were measured by fitting ellipse to their internal circle and the 3 dimensional positions of the targets were obtained through usual geometric construction, correcting for distortion of the optics. The mean radius of the joint was found to be 556.91 mm, in a zone where the joint-radius varies from 550 to 560 mm. At small deflections of the cylinder, we have compared the deflection normalised by the distance to the anchoring. The optical measurements are within 6% of the reference value given by the deflection sensor positioned on the diametrically opposed side of the target network. The elongation was also measured between targets in line with the cylinder axis, and compared with the sensor disposed in the field of view. The reference elongation curve of extent 0 to 0.25 mm- was framed by the optical measurements, which had a noise of 0.02 mm. The first results obtained with low-cost digital consumer camera appears promising, although a more sophisticated matching and following process is needed in order to clearly asses their use. Further studies are underway to match the texture of the zone in between the targets mesh, in order to increase the density of measuring points, and get better comparison with local strain gauges that are present in the target network.|
|JRC Institute:||Institute for the Protection and Security of the Citizen|
Files in This Item:
There are no files associated with this item.
Items in repository are protected by copyright, with all rights reserved, unless otherwise indicated.