Geoscience Reference
In-Depth Information
a
b
Time
Time
tt+Δt
t
t+Δt
Fig. 3 Single frame-double exposure (a) and double frames-single exposure (b) methods
influencing also the recording modes. Even if different coding sequences can be
adopted (double- or multi-exposure, see Raffael et al.
2007
), in the following only
double-exposure procedures, the actually usually adopted, are considered.
The recording modes can be separated in two main branches: single frame-
double-exposure method (Fig.
3a
), which captures the illuminated images into a
single frame; double-frames-single-exposure method (Fig.
3b
), which provides
a single illuminated image for each illumination pulse. Historically, in conjunction
with photographic cameras, a single-frame method was used, analyzing the images
using an autocorrelation process. The principal distinction between the two types
is that the single-frame method does not furnish information on the temporal order
of the illumination pulses, with an ambiguity on the direction of the velocity
vector. To overcome this ambiguity, additional costly efforts are required (i.e.,
rotating mirror, birefringent crystal, color coding, etc.). The double-frame-single-
exposure method has the advantage that the direction of flow is automatically
determined. Moreover, since it was observed that cross-correlation between two
separately recorded images is superior to the autocorrelation of double exposures,
the double frames-single exposure method is the one usually adopted in standard
PIV systems.
High speed CCD cameras are now available. The camera and the laser are
connected through a synchronizer, which is controlled by a computer and dictates
the timing of the camera sequence in conjunction with the firing of the laser (Fig.
2
).
2.2
Image Analysis, Preprocessing and Postprocessing
Procedures
The algorithm for image analysis is related to the recording procedure and to the
density of tracer particles images. Qualitatively three different situations can be
distinguished: high, medium, and low density images (Fig.
4
). In case of high image
density, it is not possible to individuate single particles, because they overlap and
form speckles. This situation is called Laser Speckle Velocimetry (LSV) and it is
not considered here. When the image density is low, individual particles can be
detected and images corresponding to the same particle can be recognized by visual
