Skip to main content
main-content

Über dieses Buch

Digital image processing has become a key technology in the area of manu­ facturing and quality control. Increasing quality demands require inspection of every single part, which in turn will lead to a much more widespread use of automatic visual inspection systems in the near future. Furthermore, the documentation requirements of ISO 9000 and similar quality control standards can only be met by fully automated, networked inspection systems. On the other hand, despite a multitude of successful applications, digital image processing has not yet established itself as an accepted element of man­ ufacturing technology. This holds true for the industrial practice as well as for the training of engineers. Digital image processing is still widely regarded as some kind of secret lore, mastered only by a small number of expensive -- experts. This impression of incomprehensibility frequently leads to the accusation of unreliability. The manufacturers of digital image processing systems in the industry are not least responsible for this state of affairs, due to their policy of giving the customer as little information as possible about the methods and technology used to inspect his products.

Inhaltsverzeichnis

Frontmatter

1. Introduction

Abstract
As we have already mentioned in the preface, increasing demands on production quality and documentation have made digital image processing one of the key technologies of modern manufacturing. Nevertheless, its industrial application is not yet commonplace, especially because of a lack of understanding for this relatively young technology. This book was written in order to remedy this condition, which was in part created by the image processing industry itself. As with all areas in which PCs are increasingly used, a trend to give the user more possibilities for application development became apparent in image processing. This makes it also necessary to equip the user with adequate know-how.
Christian Demant, Bernd Streicher-Abel, Peter Waszkewitz

2. Overview: Image Preprocessing

Abstract
Preprocessing algorithms frequently form the first processing step after capturing the image, as you will see in many of the examples in the following chapters. Therefore, we will start the overview chapters with an introduction to image preprocessing. A more comprehensive overview of preprocessing algorithms can be found in Klette and Zamperoni. 1996. To give a clear conceptual idea of the effect of the various operations we will use very simple synthetic sample images in this chapter. The application examples in the following chapters use many of these algorithms on real-world industrial images.
Christian Demant, Bernd Streicher-Abel, Peter Waszkewitz

3. Positioning

Abstract
It might appear unusual to start a detailed discussion of the various application areas of industrial image processing with positioning instead of object recognition since an object must first be found before its position can be determined. However, object recognition is a rather broad term and frequently necessitates a multitude of functions to be able to assign objects to a category. Positioning, however, is structurally speaking not necessarily algorithmically a rather simple affair as soon as the object in question has been found. The only necessary qualification is the segmentation of a reference object. The main reason to begin with this topic is that it represents an absolutely essential ”auxiliary science“.
Christian Demant, Bernd Streicher-Abel, Peter Waszkewitz

4. Overview: Segmentation

Abstract
The concept of an object is central to the solution approach in section 1.5. Indeed, it is decisive for the nature of industrial image processing since its purpose is always to gather information about objects existing in the real world represented in image scenes. In the introductory example in section 1.6 and throughout the chapter on positioning we have frequently used segmentation methods, i. e. algorithms which isolate objects from the scene. In these sections we have simply assumed that such methods exist and achieve the desired effects. This chapter will now present the most important segmentation methods among the variety of such algorithms that have developed over time.
Christian Demant, Bernd Streicher-Abel, Peter Waszkewitz

5. Mark Identification

Abstract
The term mark identification is sometimes used for the recognition of very simple markings, comparable to the marks in a multiple-choice test. We will use the term here in a broader sense for the recognition of any type of marking used to label products for identification, be it plain text. bar code, matrix codes, etc.
Christian Demant, Bernd Streicher-Abel, Peter Waszkewitz

6. Overview: Classification

Abstract
This chapter will give a brief general introduction into the field of classification and an overview of some important types of classifiers.
Christian Demant, Bernd Streicher-Abel, Peter Waszkewitz

7. Dimensional Checking

Abstract
Dimensional checking or gauging is one of the most demanding applications of digital image processing, algorithmically as well as with regard to systems engineering and facility construction. It is actually possible to reach accuracies of just a few light wavelengths, but this requires considerable effort. As in every technical discipline, precise results cannot be achieved without corresponding diligence, especially with regard to peripherals, selection of components, mechanical setup, illumination, and image capture. Quality lost in the sensory chain is lost for ever. For this reason, an overview chapter on illumination and image capturing will immediately follow this chapter.
Christian Demant, Bernd Streicher-Abel, Peter Waszkewitz

8. Overview: Image Acquisition and Illumination

Abstract
The importance of image acquisition and illumination technology for image processing cannot be overemphasized. The characteristics of all parts of the sensoric chain — starting with the lens and the light-sensitive elements and leading to the transmission to the computer — affect the quality of the digital image. Hence, they are decisive factors for the quality of the results that can be obtained from the image. Quality loss incurred here can hardly be compensated for by software technology. All application areas of digital image processing depend on high quality images, but none more so than optical gauging, because of the high demands on accuracy in this area. Therefore, we will pay particular attention to image distortions that make it impossible to obtain accurate measurements.
Christian Demant, Bernd Streicher-Abel, Peter Waszkewitz

9. Presence Verification

Abstract
Viewed from the perspective of the problem to be solved, presence verification appears easily grasped and structurally rather simple. This is the reason why presence verification is often used as an introductory example to image processing. In reality, the concept of presence verification is quite difficult to define precisely: is it simply counting objects or are properties of these objects of importance? Are these properties simple features or is the entire appearance of the object to be considered? Do properties relate to individual objects only or do we have to take relationships between objects into account, for example when doing assembly verification.
Christian Demant, Bernd Streicher-Abel, Peter Waszkewitz

10. Overview: Object Features

Abstract
Most examples in the preceding chapters used various features to check image objects for validity. It is time to present an overview of such features selected from the immense variety described in the image processing literature on the basis of their practical usefulness in industrial applications. At the same time, we will present some further insights into the difficulties encountered when applying common everyday notions to the discrete world of digital images.
Christian Demant, Bernd Streicher-Abel, Peter Waszkewitz

11. Outlook: Visual Inspection Projects

Abstract
In the preceding chapters we have described a variety of methods and algorithms from the field of image processing and shown their application in various industrial projects. In this chapter, we will conclude this introduction to optical quality control with some important aspects regarding the implementation of industrial image processing projects. Some of these correspond to general rules for industrial project management: others are specific to image processing.
Christian Demant, Bernd Streicher-Abel, Peter Waszkewitz

Backmatter

Weitere Informationen