System and method for remote full field three-dimensional displacement and strain measurements
Inventors
Michopoulos, John G. • Iliopoulos, Athanasios • Andrianopoulos, Nikos P.
Assignees
US GOVERNMENT IN NAME OF SECRETARY OF NAVY • US Department of Navy
Publication Number
US-9046353-B2
Publication Date
2015-06-02
Expiration Date
2032-08-02
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Abstract
A method and system for measuring full field deformation characteristics in three dimensions of a body upon which a pattern of visible marks has been applied. The method includes receiving images of the pattern of marks from at least two digital video cameras as the specimen deforms. A computer processor identifies the centroids of each of the two camera images of the body in a first frame, matches the centroids of the marks in the two images, and generates a three-dimensional representation of the centroids based on the two images, and repeats these steps for the images from the two cameras in a subsequent frame. The computer processor calculates the displacement vector between a three dimensional representation of the centroids and the subsequent three dimensional representation of the centroids, and calculates full field displacement and strain fields based on the displacement vector.
Core Innovation
The invention is a computer-implemented method and apparatus for measuring full field deformation characteristics in three dimensions of a body upon which a pattern of visible marks (dots) has been applied. The method involves receiving images of the pattern from at least two cameras from different angles, identifying the centroids of the marks in images from both cameras in sequential frames, matching corresponding centroids between images and frames, and generating three-dimensional representations from the matched centroids. The displacement vectors between the three-dimensional centroid positions before and after deformation are then calculated, and used to compute full field displacement and strain fields using mesh-free approximation techniques.
The invention addresses the problem in existing digital image correlation and pure grid methods, which are sensitive to the application of uniform grids or speckle patterns and have difficulties with irregular shapes and accurate point matching. Prior methods suffer from challenges related to the application and detection of deformation patterns, the need for specialized grids, and limitations in tracking full three-dimensional deformation, especially on non-planar or irregular surfaces. The invention overcomes these by using a random dot pattern with mesh-free surface approximations and computer vision techniques to reconstruct 3D displacement and strain fields accurately and efficiently.
The solution includes preparing a deformable specimen by applying a random distribution of optically distinct spots, capturing images before and after deformation using two or more calibrated cameras, processing the images to find centroids of the spots via light integration and segmentation algorithms, matching centroids between stereo images and between time frames, reconstructing 3D point clouds of the centroid positions, and interpolating displacement and strain fields with mesh-free methods. The system includes software (ReMDiS-3D) with a graphical user interface for calibration, segmentation, geometry definition, analysis, and visualization, enabling real-time processing and display of displacement and strain with color-coded fields.
Claims Coverage
The claims describe multiple inventive features related to a method and an apparatus for measuring three-dimensional deformation using optical patterns and computational processing. There are two independent claims: one directed to the computer-implemented method, and another directed to an apparatus configuration.
Method for three-dimensional full field deformation measurement using visible dot patterns
A computer-implemented method including: receiving sequences of images from at least two imaging devices oriented at different angles; identifying centroids of the dots in images from each device; matching centroids between images in a frame; generating three-dimensional representations of the centroid patterns; matching three-dimensional centroid representations between frames; calculating displacement vectors between frames; and computing full field three-dimensional displacement based on the displacements.
Calculating full field strain tensor from displacement vectors
Extending the method by calculating the full field strain tensor based on the computed displacement vectors of the centroids to obtain strain distributions over the deforming body.
Identification and filtering of centroids via light integration and segmentation
Using light intensity integration to find centroids of dots and eliminating irregular or irrelevant objects from images based on criteria including intensity threshold, pixel area range, aspect ratio, moment of inertia, major axis direction, and compactness ratio to improve centroid detection and accuracy.
Selection of mesh free approximation parameters based on user input
Receiving performance characteristics from a user through a graphical user interface and selecting mesh free approximation parameters accordingly for optimized calculation of displacement and strain fields.
Displaying and storing graphical representations of displacement and strain data
Graphically presenting and storing acquired images, user inputs, and calculated displacement and strain components using color scales, and plotting displacement, elongation, or strain versus time or image frames to facilitate visualization and analysis.
Apparatus for three-dimensional deformation measurement with stereo cameras and processor
An apparatus comprising at least two cameras facing the deformable body at different angles to acquire image sequences, and a computer processor programmed to identify centroids in images, match centroids between images and frames, generate three-dimensional centroid representations, calculate displacement vectors between frames, and calculate full field displacement and strain fields using mesh free approximation.
The independent claims collectively cover an innovative method and apparatus that employ stereo imaging of a visible dot pattern on a deforming body, computer vision-based identification and matching of centroid points, three-dimensional reconstruction of centroid positions, and mesh-free approximation techniques to compute full three-dimensional displacement and strain fields, with provisions for filtering data and user-controlled parameters for performance optimization and visualization.
Stated Advantages
Provides efficient and inexpensive full field displacement and strain determination in three dimensions over deforming bodies of various shapes.
Enables accurate measurement of out-of-plane deformations and strain fields using non-contact optical methods.
Utilizes mesh-free approximation methods capable of handling irregular domains and randomly distributed points, ensuring high order continuity and flexibility.
Offers real-time visualization, analysis, and user interactivity via a comprehensive graphical user interface.
Improves accuracy and reliability of centroid detection by filtering irregular objects based on multiple image feature criteria.
Documented Applications
Characterization of material deformation behavior, including composite material specimens subjected to mechanical loading resulting in in-plane and out-of-plane displacements and strains.
Full field strain and displacement measurement for experimental mechanics and materials testing, including tests conducted using mechatronically automated robotic machines such as six degree-of-freedom hexapod test machines.
Real-time monitoring and analysis of deforming specimens during mechanical tests, involving tension and combined loading conditions up to fracture.
Interested in licensing this patent?