Previous: SAMPLEX Next: How to get started

What Is SAMPLEX

SAMPLEX is the X version of the color classification tools developed in the Agricultural Engineering Department at Purdue University by Dr. C.J. Précetti and Dr. G.W. Krutz.

Color classification is the separation of images into different colors, for analysis purposes. Generally objects are separated into color groups describing a physical property (red ripe apple, yellow unripe apple, brown over-ripe apple...).

The following example presents a color classification with SAMPLEX of seed corn. This image was taken at 1/2000 th second on a conveyor belt, the classification objective is to estimate the amount of husk delivered with corn. The color classes are:

Figure 0: SAMPLEX example on seed corn

SAMPLEX is meant to replace software developed this last three years for color classification. It is also designed for undergraduate labs, using the wokstations (32 Sun) available in the department. The objective is to familliarize the futur Food Process engineer with color machine vision for product and system control.

It regroups:

SAMPLEX is designed for 8-bit displays and 256 colors.
(Tutorial on creating and using colormaps : discover what is a colormap and how it influences the way images look.)

The classification process has three steps:

  1. Sampling picture elements (pixels) for each color class
  2. Creating a classifier using statistical models
  3. Classifying the image with the classifier
Classifiers can be created using many statistical and artificial intelligence techniques. In SAMPLEX the classification algorithms are linear and quadratic Bayes classifiers from the software SPR, and the neural network BLC2 classifier from nSPR.

SAMPLEX does not allow every color classification process available in SPR, nSPR and Purclass. A description of SPR and other classification software can be found in C.J.Précetti Ph.D. thesis at Purdue University.

Previous: SAMPLEX Next: How to get started