From photogrammetry, computer vision to structural response measurement

Measuring displacement response of large-scale structures is necessary to assess their performance and health condition. Current global position systems (GPS), with a significant cost, can provide point coordinate measurement with an accuracy of ±1cm horizontally and ±2cm vertically at a sampling frequency up to 20 Hz. Photogrammetry is a measurement technology in which the three-dimensional coordinates of points on an object are determined by measurements made with two or more photographic images taken from different positions. This technology is regaining its popularity in various engineering disciplines due to a recent remarkable evolution in the consumer digital cameras. The image resolution of these cameras has increased rapidly from below 1 million pixels a few years ago to over 10 million pixels today, with little increase or even decrease in cost. Image sequences recorded by these cameras contain both spatial and temporal information of the target object; hence can be used to extract the object's dynamic characteristics such as natural frequencies and mode shapes. Research indicates that accuracy in the order of as high as 1 part in 30,000 can be achieved when cameras are properly calibrated in the field and multiple high-resolution images are used. At this level of accuracy, point positions of a 30m object would be accurate to 1mm at 68% probability and responses of large-scale structures can be measured for further meaningful processing and interpretation. In this paper, several image-based sensing techniques that can be used for measuring response of large-scale structures are presented. These techniques are developed based on some methods from photogrammetry and computer vision such as the optical flow method and the egomotion estimation. Examples used to illustrate these techniques include response measurement of buildings and cable-stayed bridges. Results show that the image-based sensing techniques have a great potential for accurately measuring displacement responses of these large-scale structures.

[1]  David V. Jáuregui,et al.  Noncontact Photogrammetric Measurement of Vertical Bridge Deflection , 2003 .

[2]  Bernhard P. Wrobel,et al.  Multiple View Geometry in Computer Vision , 2001 .

[3]  Peter Meer,et al.  Balanced Recovery of 3D Structure and Camera Motion from Uncalibrated Image Sequences , 2002, ECCV.

[4]  M. Isard,et al.  Automatic Camera Calibration from a Single Manhattan Image , 2002, ECCV.

[5]  B.-Z. Yuan,et al.  Using vision technique for the bridge deformation detection , 1988, ICASSP-88., International Conference on Acoustics, Speech, and Signal Processing.

[6]  D. D. Lichti,et al.  MEASUREMENT OF DEFLECTIONS IN CONCRETE BEAMS BY CLOSE- RANGE DIGITAL PHOTOGRAMMETRY , 2002 .

[7]  Y. B. Aw,et al.  PHOTOTRIANGULATION USING CCD CAMERA IN CLOSE-RANGE ENVIRONMENT , 1993 .

[8]  Carlo Tomasi,et al.  Comparison of approaches to egomotion computation , 1996, Proceedings CVPR IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[9]  Jianwen Liang,et al.  Digital image processing for non-linear system identification , 2004 .

[10]  Hans-Gerd Maas,et al.  Automatic deformation measurement with a digital still video camera , 1997 .

[11]  Piotr Olaszek,et al.  Investigation of the dynamic characteristic of bridge structures using a computer vision method , 1999 .

[12]  C. Fraser,et al.  Monitoring the thermal deformation of steel beams via vision metrology , 2000 .

[13]  S. Patsias,et al.  Damage Detection Using Optical Measurements and Wavelets , 2002 .

[14]  Fred B Bales,et al.  CLOSE-RANGE PHOTOGRAMMETRY FOR BRIDGE MEASUREMENT , 1984 .

[15]  Mubarak Shah,et al.  Review of computer vision education , 2003, IEEE Trans. Educ..

[16]  K. Torlegård The International Society for Photogrammetry and Remote Sensing , 1987 .