C. R. Coggrave

Real-time visualisation of deformation fields using speckle interferometry and temporal phase unwrapping

Abstract A real-time system for analysing data from speckle interferometers, and speckle shearing interferometers, has been developed. Interferograms are continuously recorded by a digital camera at a rate of 60 frames s −1 with temporal phase shifting carried out at the same rate. The images are analysed using a pipeline image processor. With a standard 4-frame phase-shifting algorithm (phase steps of π /2), wrapped phase maps are calculated and displayed at 15 frames s −1 . These are unwrapped using a temporal phase unwrapping algorithm to provide a real-time colour-coded display of the relevant displacement component. Each camera pixel (or cluster of pixels) behaves in effect as an independent displacement sensor. The reference speckle interferogram is updated automatically at regular user-defined intervals, allowing arbitrarily large deformations to be measured and errors due to speckle decorrelation to be minimised. The system has been applied to the problem of detecting sub-surface delamination cracks in carbon fibre composite panels.

Progress in phase unwrapping

This paper reviews some of the main developments in numerical phase unwrapping techniques that have been introduced over the past five years. In quality control applications of wholefield optical techniques it is often necessary for the phase unwrapping to be both robust and performed within a guaranteed time period of a few seconds or less. Deterministic algorithms that have data-independent run times are therefore desirable. The temporal phase unwrapping algorithm has been implemented on a pipeline image processor which is able to calculate phase maps from phase-shifted fringe patterns acquired at 30 frames s-1, and to unwrap the resulting phase maps in real time. One application of the system is measurement of surface profile. The fringe patterns are projected at 30 frames s-1 using a high resolution data projector. The spatial frequency is reduced exponentially from the maximum value. The sequence of phase values at a given pixel is then unwrapped independently of the other image pixels, and all the intermediate phase values contribute in a least squares sense to the final range estimate for the pixel. A total acquisition and processing time of 0.7 s has been achieved for a maximum spatial frequency of 16 fringes across the field of view.

Combining digital image correlation and projected fringe techniques on a multi-camera multi-projector platform

This paper presents how a shape measurement system (SMS) based on projected fringes can be combined with a 2-D digital image correlation (DIC) technique to accurately measure surface profile and 3-D displacement fields at the same time. Unlike traditional 3-D DIC techniques, the proposed method can measure discontinuous surfaces as easily as smooth ones. The method can also be extended to a multi-camera multi-projector system, and thus complete 360° 3-D displacement fields can be obtained within a single global coordinate system. Details of the algorithm are presented together with experimental results.

High Speed Measurement of Discontinuous Surface Profiles

The measurement of surface profile by projecting phase-stepped fringe patterns at an angle to the observation direction is well known. Ambiguous range data results, however, when the object has discontinuities in its profile. The ambiguities can be prevented by projecting fringes of varying spatial frequency. In this paper we describe an approach which combines high accuracy and reliability. The spatial frequency is reduced exponentially from the maximum value. The sequence of phase values at a given pixel is then unwrapped independently of the other image pixels, and all the intermediate phase values contribute in a least squares sense to the final range estimate for the pixel. The algorithm has been implemented on a pipeline image processing system. The fringe patterns are projected at 30 frames s−1 using a high resolution data projector. Images are acquired and analysed in real time, at the same framing rate. A total acquisition and processing time of 0.75 s has been achieved for a maximum spatial frequency of 16 fringes across the field of view.

Optimization of a shape measurement system based on spatial light modulators

A high speed shape measurement system has been developed based on the method of projected fringes. An optimized sequence of fringe pitches is used in which the spatial frequency is reduced exponentially from the maximum value. Temporal unwrapping of the phase values at each fringe pitch provides an independent co-ordinate from each camera pixel. A commercial data projector produces the fringe patterns at a rate of 30 s-1; the images are analyzed in real time by means of a pipeline image processor. A total acquisition and analysis time of 0.87 s is required for 250,000 coordinates. The main results of an experimental study to optimize the measurement precision of the system (from the original value of one part in 2,000) are also described. Issues that have been investigated include (1) the relative performance of different phase-shifting algorithms (4-frame, 7-frame and 15-frame); (2) the relative performance of different temporal phase unwrapping algorithms; and (3) the effect of projector defocus. At the optimum focal position, a measurement precision of better than one part in 20,000 of the field of view was achieved.

Real-time speckle interferometry fringe analysis system

A real-time system for analyzing data from speckle interferometers, and speckle shearing interferometers, is presented. Interferograms are recorded by a CCD camera at a rate of 60 frames s-1, with temporal phase shifting carried out at the same rate by means of a PZT-mounted prism in the reference beam. The images are analyzed using a pipeline image processor consisting of two Datacube MaxVideo 250 VME boards. With a standard 4-frame phase-shifting algorithm, wrapped phase maps are calculated at 15 frames s-1; these are then unwrapped in real time using a temporal phase unwrapping algorithm. The main advantage of temporal over spatial unwrapping methods is that unwrapping errors due to noise and specimen boundaries do not propagate spatially. Pseudo-color maps of displacement distribution are displayed throughout the measurement period in real time. The paper is illustrated by application of the technique to the rigid body motion of a flat plate under controlled conditions.