4.1. Implementation of a CVS
The CVS implemented in this research has principally two parts: (i) the image acquisition set up as showed in
Fig. 2; (ii) the digital image processing part is composed of a set routines to pre-process the acquired and stored
images, to segment the images and separate objects of interest from the background, and to convert RGB images
of the objects of interest into L*a*b* units of color. In this research we focus on part (ii) integrating the digital image
analysis routines programmed in Matlab code for pre-processing, segmentation and color conversion in order to
develop a computer program and make easier and faster color determination in L*a*b* units. Digital RGB images
were captured with a previously implemented image acquisition
system and stored in a PC for future processing
(Pedreschi et al., 2004).
The Matlab commands used in this system are
explained in Table 1. The color image is firstly stored in
a matrix I. In order to reduce the computational time of
processing the images can be sub-sampled. If the images
are noisy they can be filtered using a low pass filter (Mathworks,
2005). Afterwards, the RGB image is segmented
using the algorithm outlined in Mery and Pedreschi
(2005). In this step, the image of the potato chip is separated
from the background producing a binary image R
(Fig. 3). Finally, the RGB image is converted into a