Filippo Bertolino

On the performance of some unwrapping algorithms

This work attempts an objective comparison in terms of performance and execution speed of some of the best known phase unwrapping algorithms. First the algorithms chosen, grouped into 4 classes (sequential methods, residues methods, global least square integration methods, others) are described. Then the influence of the weighting function on each of them is investigated. So as to obtain quantitative results it is necessary to use synthetic images, so that the exact solution is known, the technique for generating these images is also illustrated. Lastly, the algorithms performance in terms of influence of the weight functions, robustness and execution time are discussed.

A posteriori compensation of the systematic error due to polynomial interpolation in digital image correlation

Abstract. It is well known that displacement components estimated using digital image correlation are affected by a systematic error due to the polynomial interpolation required by the numerical algorithm. The magnitude of bias depends on the characteristics of the speckle pattern (i.e., the frequency content of the image), on the fractional part of displacements and on the type of polynomial used for intensity interpolation. In literature, B-Spline polynomials are pointed out as being able to introduce the smaller errors, whereas bilinear and cubic interpolants generally give the worst results. However, the small bias of B-Spline polynomials is partially counterbalanced by a somewhat larger execution time. We will try to improve the accuracy of lower order polynomials by a posteriori correcting their results so as to obtain a faster and more accurate analysis.

A Low-Cost Residual Stress Measuring Instrument

The hole-drilling method is the approach most used for residual stress measurement and most of the commercially available instruments are based on this working principle. The idea is quite simple: drill a small hole on the surface of the components and measure the strain/displacement components resulting on the surface. The stress components can be estimated from these data with a reverse calibration process. The measurement of the displacement/strain field is usually performed using either strain gauges or interferometric optical methods, thus the cost of the instruments is significant.

Systematic errors due to polynomial interpolation in digital image correlation: analysis of some interpolating kernels and experimental validation

The accuracy of image correlation algorithms is severely affected by errors introduced by the interpolation step. This article surveys some of the most common interpolating and approximating kernels, focusing on the various sources of error-signal attenuation, phase errors and frequency spectrum-to identify the best performing algorithms. The results of the analysis are experimentally validated by comparing the Digital Image Correlation estimated displacements with the interferometric estimated values.

Comparing Two Different Approaches to the Identification of the Plastic Parameters of Metals in Post-necking Regime

In the past 20 years the growing computation power availability encouraged experimental mechanics specialists to couple full field measurements with FE methods to raise the so called hybrid experimental-numerical methods. A typical example is the identification of the plastic parameters of metals starting from experimental data.

Semiautomatic approach of grating techniques

Abstract After a brief description of the experimental set-up used for the strain analysis of composite laminates with graphite fibres and thermoplastic matrix, the results obtained using a software code suitable for reading and analysing fringe patterns are presented. In particular the program uses two different approaches: the first semi-automatic with algorithms proper of fringe tracking and the second based on Fourier transform.

Integrated digital image correlation for residual stress measurement

Residual stress measuring by optical methods has been proposed by several authors in the past years; nevertheless the technique is still confined to optical laboratories: indeed the large sensitivity of the optical methods allows measuring very low stress values with high reliability, but these advantages are counter-balanced by the high sensitivity to vibrations, which makes very difficult performing the measurements outside of optical laboratories. Digital Image Correlation (DIC) is a possible alternative: indeed this technique is much less affected by vibrations, but its sensitivity is quite low, thus negatively affecting the accuracy of results. This work proposes a variant of Digital Image Correlation, known as iDIC (integrated DIC), to perform residual stress measurement. Since this approach directly integrates in the formulation the hole drilling shape functions, it overcome most of the problem of standard DIC; in this way it is possible to obtain accurate results without using interferometric techniques.

Phase Unwrapping Algorithms: A Comparison

Over the last few years a great deal of work has been done to improve the reliability of phase unwrapping algorithms. Many articles are to be found in the technical literature claiming the robustness and noise immunity of the proposed algorithm.

An Application of White Light Profilometry Using Geometric Phase Shifting

Laser surface action profilometry is a well known technique for measuring the statistical parameters of a profile. However it is unable to deal with surfaces having discontinuities larger than half the laser wavelength. This limitation can be overcome using either two laser sources having very close wavelength or a single white light source. Both techniques present some problems: the first requires two laser sources or a variable wavelength laser and involves a numerical only solution. On the other hand, the white light system is based on a physical property (the short coherency length of the source) but requires an achromatic shifter. We have evaluated the accuracy of the white light system taking into account the results of the two-laser reference configuration. In order to achieve this objective an experimental set-up has been devised that can acquire data on the same area. Phase shifting has been performed using an achromatic geometric shifter in both cases, while two different procedures have been followed to process the fringe images. In conclusion, the results obtained with white light are very close to those obtained with the two-laser system. The main drawback is the very large amount of data needed to perform image processing.

A Meshless Global DIC Approach

The formulation of global Digital Image Correlation (DIC) is almost identical to standard (i.e. subset-based) DIC: indeed, in both cases the solution algorithm tries to minimize the error between the reference and test image by adjusting the shape-function-controlling parameters, but in the former, a modification of the parameters affects the displacement field globally, whereas in standard DIC the modification is local.