Important Parameters for Image Color Analysis: An Overview - Visual Computing: Scientific Visualization and Imaging Systems - page 88

study varying the background color in the selection of pomegranate seeds, con-
cluding that the blue background contributed to a better segmentation.
The diversity of industrial products (which have a variety of sizes, shapes,
textures, and color) emphasizes the importance of defining standardized parame-
ters during the classification and packing [ 1 ], since inadequate illumination may
hinder the identification of defects and the color of the products analyzed [ 9 ] and
the lack of standardization of the background color used for the classification of
products (usually conveyor belts or benches) can impair selection process.
4.5 Case Study
Gomes et al. [ 19 ] developed a research in order to define a new method for
nondestructive testing using colorimetric techniques and computer vision for
characterizing color using digital images applied to integrated fruit production,
focusing on standardization of measurements, considering the factors of influence.
In the initial study, the banana was chosen as a case study in the development of
this research, and as a result, a new methodology was developed to characterize
the stages of maturation of banana using colorimetric analysis, proposing a stan-
dard for the industry [ 10 ].
Figure 4.5 shows an example of visual assessment in monitoring of the banana
ripening for selection and fruit trade [ 10 ] using a halogen lamp as source
(CCT = 2,856 K, CRI = 99.8) and black background.
Figure 4.6 shows the difference of perception in an image when using different
types of illumination. Figure 4.7 shows the difference in perception of the image
when different background colors are used [ 10 ].
Large color differences were found when comparing the same ripening subclass
using different sources. Such differences justify the need for greater concern about
the lighting system employed in the classification area. When comparing different
background colors, red and blue backgrounds offered most influence on the
evaluation of ripening subclasses.
Therefore, one should evaluate the characteristics of interest to define the
parameters to be used in image acquisition in order to get a better image and better
accuracy in measuring color from the image. For improving the results it is sug-
gested the following steps for defining an image acquisition system for color
analysis: a study of the sample; a study of the best background color; a study of the
best light source; and finally the calibration of the measurement system under
these conditions (Fig. 4.8 ).