**1. Introduction**

In the last years many industries have increased the exploitation of vision systems applied to several fields. This fact is basically due to the technological progress that has been reached by these systems: the reliability of vision systems allows the industries to achieve considerable cost savings in terms of both material and human resources.

Among the most important applications of vision systems in the industrial field there are robot positioning and driving, code reading, non contact measuring as well as quality control and monitoring.

In particular quality controls are nowadays performed through vision systems in many industries; these systems guarantee reliability comparable and sometimes greater with respect to human operators, especially for a large quantity of products. In fact, vision-based automatic inspection systems allow to process a huge amount of data in a very small time with respect to human performance.

The vision systems are usually composed by several components: a set of cameras to capture the images, an illumination system and a system for the image processing and classification. These systems are able to capture images of a wide range of products in order to find defects which do not fit the quality standards of the considered industrial production process.

The present chapter deals with the application of Fuzzy Inference Systems (FIS) for the classification of images, once they have been pre-processed through suitable algorithms. One of the main reasons for exploiting a FIS to this aim lies in the possibility of approaching the problem in a way similar to human reasoning. In fact a FIS allows:


Fuzzy Inference Systems Applied to Image Classification in the Industrial Field 245

video camera, in order to plan future operations. The automatic systems which carry out visual inspection by means of machine vision are often called *Automatic Visual Inspection* 

Artificial vision is defined as the set of computational techniques which aim to create an approximate model of the three-dimensional world from two-dimensional projections of it. A classic problem concerning artificial vision is to determine whether there are specific objects or activities in the image. Another problem, which is solved by vision systems, is to reconstruct a three dimensional model about the scene to be analysed, given one or more two-dimensional images of it (Klette et al., 1998). The optical techniques which are widely adopted can be divided in two approaches: active methods and passive methods. Active methods regard the case where the scene is radiated by appropriate electromagnetic radiation, while passive methods regard the case where the images of the scene are the real images. Active methods are more expensive than passive ones and they are not always applicable. In contrast, passive methods have lower resolutions than the active ones. Both approaches adopted the visual cues that are used by human vision system to retrieve the depth of the scene projection on the retina, such as blurring and

There are many applications of artificial vision in industrial field such as defect detection, robot placement, robot orientation and robot guidance (Lowe & Little, 2005), codes reading, classification. In the last decades the Artificial Vision has evolved into a mature science, which embraces different markets and applications becoming a vital component of

*Systems* (Jarvis, 1979).

Fig. 1. Scheme of a simple industrial vision system.

other optical phenomena.

advanced industrial systems (Lanzetta, 1998).

Moreover, FIS-based classifiers provide the further advantage of a great flexibility, thanks to the possibility to add rules without affecting the remaining parts of the inference machine.

Clearly the preliminary image processing phase should be developed with the aim to extract the most suitable features from the images to be fed as input to the fuzzy system.

The chapter will be organised as follows: a preliminary generic description of the main features of FIS-based image classifiers will be provided, with a particular focus on those general applications where a preliminary ad-hoc image features extraction phase is included. The advantages and limitations of FIS-based classifiers toward other algorithms will be presented and discussed. Afterwards, some case studies will be proposed, where FIS-based image classifiers are applied in an industrial context.
