Characterization of Complex Patterns: Application to Colorimetric Arrays and Vertical Structures

Introduction

A major challenge of a typical machine inspection process is to provide rugged and cost-effective solutions for real-time problems. Such systems use different and appropriate sensors for the automatic detection and identification of various suspicious components, as production defects or VOSs and serve as a valuable process feedback and control utility. A step towards the cost reduction of such processes is, to define approaches that can be applied to a wide range of applications, provided the efforts to adapt such a solution to a specific task are minimal.

The major perspective of this chapter is to propose a general approach for real-time pattern interpretation for different quality control processes. The case of structured patterns for industrial specular surface inspection will serve as basis. The method using the interpretation of a basic periodical and vertical light pattern was recently proposed for the inspection and characterization purposes of cylindrical specular surfaces (Caulier et al., 2007; Caulier et al., 2009). This sensor technique permits the visual enhancement and discrimination of various defective parts of the specular surfaces similar to colorimetric sensors allowing the visual enhancement of organic components. Both methods use the interpretation of the visual information to discriminate between different types of components, whether these are defective metallic surfaces or organic and volatile.

In order to propose a new real-time color array description, a transformation function between color arrays and structured images is proposed. This transformation is based on a different representation of each important feature of both arrays, i.e. the hue, saturation, and value components for the former and the intensity, left and right deviation for the latter. This chapter is therefore dedicated to the generalization of periodical and vertical structured interpretations. The adaptation to color array description is then straightforward.

Various pattern analysis techniques have been proposed since the mid 1960s, i.e. since computers were able to solve information handling problems. According to (Raudys & Jain, 1991), the main steps defining a typical pattern recognition process are the following: data collection, pattern class's formation, characteristic features extraction, classification algorithm specification and estimation of the classification error. The results of each of these steps can be used in a feedback procedure for optimization of the final result. Then, depending on the size and the representativeness of the reference data set, various classification methodologies can be applied (Witten & Eibe, 2008).

With the interpretation of such stripe patterns, the task consists of the computation of the optimal pattern analysis methods in terms of highest classification rates. The core of our approach is dedicated to the retrieval and selection of the most appropriate feature sets for the characterization of vertical stripe structures.

For the purpose of optimizing the retrieval of the most appropriate features for stripe patterns characterization, a three-step method is proposed. Indeed, such hierarchical feature selection procedures are considered to be particularly suited to complex content-based image description tasks as we have here (Dy et al., 2003; Peng et al., 2005). Furthermore, in order to address the stripe classification task in general, and in accordance to the recommendations of (Raudys & Jain, 1991), we emphasize the fact, that several feature extraction and feature selection methods, but also various classification algorithms and classification methodologies, are taken into consideration.