Zusammenfassung
The standard method of shrinkage measurement consists of
immersion of the product in a fluid in order to
calculate the volume changes before and after
drying. It is destructive and time-consuming and
also is not a practical method to be used in online
drying monitoring systems. To date, it has been
tried for measuring shrinkage based on passive
stereo vision. But no report has been provided so
far on the accuracy of this technique and its
comparison with conventional method of measuring
volumetric shrinkage. On the other hand, because of
the small size of dried foodstuff products, it does
not seem that the stereo vision to be able to
extract high detail point clouds from the surface of
objects. Therefore, this research was conducted in
order to study the potential use of 3D laser
scanning for measurement of the volumetric shrinkage
of some horticultural products during drying
process. To this end, a calibrated 3D laser imaging
system was applied in order to precisely scan the
surface of some small size horticultural products
(including plum, fig, date, and button mushroom),
which take non-symmetric form during the drying
process. 2D image of samples was also taken to
predict the volumetric shrinkage by various texture
analysis methods. Drying was carried out by a
convective dryer. The results indicated a
significant superiority of 3D laser imaging compared
to 2D imaging. The value of correlation coefficient
and mean absolute percentage error of multilayer
perceptron artificial neural networks models created
based on selected spatial features of point clouds
in predicting volumetric shrinkage for plum, fig,
date, and mushroom was obtained 0.90 and 19.48, 0.95
and 14.25, 0.78 and 23.54, and 0.87 and 9.47,
respectively.
Nutzer