from 883 556 to 300 300 pixels. Then the backgrounds of
the images were removed. Color and textural features of objects
such as contrast, entropy, nth moment and homogeneity
were extracted by image processing toolbox of the Matlab
software. Then each parameter was analyzed by statistical
methods in the SPSS software. At the end, those parameters
which were significantly different were selected to accomplish
the sorting task. Among the investigated features, the color
feature was significantly different from other parameters; so
an index based on color components i.e. Red, Green and Blue
was defined for detection algorithm. In order to compute the
coefficients of each component of an index, the Taxonomy
method was used. Taxonomy is a classification method on
the basis of a multivariate analysis of observable differences
and similarities between taxonomic groups. Classifications
based on numerical taxonomy reflect degrees of evolutionary
relationships.
During online experiments, 100 samples of fruits for each
maturity stage were selected at random. Samples were fed onto
a conveyor belt in a row. The speed of the conveyor belt was
not controlled by a micro controller. There was a sensor before
the illumination box; whenever the fruit passed the sensor, just
after a pause, a snapshot was taken and the real time captured
image was recalled to the workspace for processing. The size of
the images was reduced to 300 300 pixels. Since the inverter