
Machine Vision Algorithms and Applications
Beschreibung
Alles über E-Books | Antworten auf Fragen rund um E-Books, Kopierschutz und Dateiformate finden Sie in unserem Info- & Hilfebereich.
Weitere Details
Weitere Ausgaben
Andere Ausgaben

Personen
Markus Ulrich studied Geodesy and Remote Sensing at the Technical University of Munich (TUM) and received his PhD degree from TUM in 2003. In 2003, he joined MVTec?s Research and Development department as a software engineer and became head of the research team in 2008. He has authored and co-authored scientific publications in the fields of photogrammetry and machine vision. Markus Ulrich is also a guest lecturer at TUM, where he teaches close-range photogrammetry. In 2017, he was appointed a Privatdozent (lecturer) at the Karlsruhe Institute of Technology (KIT) for the field of machine vision.
Christian Wiedemann studied Geodesy and Remote Sensing at the Technical University of Munich (TUM) and received his PhD degree from TUM in 2001. He has authored and co-authored more than 40 scientific publications in the fields of photogrammetry, remote sensing, and machine vision. In 2003, he joined MVTec's Research and Development department as a software engineer. Since 2008, he has held different leading positions at MVTec.
Inhalt
IMAGE ACQUISITION
Illumination
Lenses
Cameras
Camera-Computer Interfaces
MACHINE VISION ALGORITHMS
Fundamental Data Structures
Image Enhancement
Geometric Transformations
Image Segmentation
Feature Extraction
Morphology
Edge Extraction
Segmentation and Fitting of Geometric Primitives
Camera Calibration
Stereo Reconstruction
Template Matching
Optical Character Recognition
MACHINE VISION APPLICATIONS
Wafer Dicing
Reading of Serial Numbers
Inspection of Saw Blades
Print Inspection
Inspection of Ball Grid Arrays
Surface Inspection
Measuring of Spark Plugs
Molding Flash Detection
Inspection of Punched Sheets
3D Plane Reconstruction with Stereo
Pose Verification and Resistors
Classification of Non-Woven Fabrics
2
Image Acquisition
In this chapter, we will take a look at the hardware components that are involved in obtaining an image of the scene we want to analyze with the algorithms presented in Chapter 3. Illumination makes the essential features of an object visible. Lenses produce a sharp image on the sensor. The sensor converts the image into a video signal. Finally, camera-computer interfaces (frame grabbers, bus systems like USB, or network interfaces like Ethernet) accept the video signal and convert it into an image in the computer's memory.
2.1 Illumination
The goal of illumination in machine vision is to make the important features of the object visible and to suppress undesired features of the object. To do so, we must consider how the light interacts with the object. One important aspect is the spectral composition of the light and the object. We can use, for example, monochromatic light on colored objects to enhance the contrast of the desired object features. Furthermore, the direction from which we illuminate the object can be used to enhance the visibility of features. We will examine these aspects in this section.
2.1.1 Electromagnetic Radiation
Light is electromagnetic radiation of a certain range of wavelengths, as shown in Table 2.1. The range of wavelengths visible for humans is 380-780 nm. Electromagnetic radiation with shorter wavelengths is called ultraviolet (UV) radiation. Electromagnetic radiation with even shorter wavelengths consists of X-rays and gamma rays. Electromagnetic radiation with longer wavelengths than the visible range is called infrared (IR) radiation. Electromagnetic radiation with even longer wavelengths consists of microwaves and radio waves.
Monochromatic light is characterized by its wavelength ?. If light is composed of a range of wavelengths, it is often compared to the spectrum of light emitted by a black body. A black body is an object that absorbs all electromagnetic radiation that falls onto it and thus serves as an ideal source of purely thermal radiation. Therefore, the light spectrum of a black body is directly related to its temperature. The spectral radiance of a black body is given by Planck's law (Planck, 1901; Wyszecki and Stiles, 1982):
(2.1)Table 2.1 The electromagnetic spectrum relevant for optics and photonics. The names of the ranges for IR and UV radiation correspond to ISO 20473:2007. The names of the colors for visible radiation (light) are due to Lee (2005).
Range Name Abbreviation Wavelength ? Ultraviolet Extreme UVVacuum UV
Deep UV
Mid UV
Near UV -
UV-C
UV-B
UV-A 1 nm-100 nm
100 nm-190 nm
190 nm-280 nm
280 nm-315 nm
315 nm-380 nm Visible Blue-purple
Blue
Green-blue
Blue-green
Green
Yellow-green
Yellow
Orange
Red
Red-purple 380 nm-430 nm
430 nm-480 nm
480 nm-490 nm
490 nm-510 nm
510 nm-530 nm
530 nm-570 nm
570 nm-580 nm
580 nm-600 nm
600 nm-720 nm
720 nm-780 nm Infrared Near IR
Mid IR
Far IR IR-A
IR-B
IR-C 780 nm-1.4 µm
1.4 µm-3 µm
3 µm-50 µm
50 µm-1 mm
Here, c = 2.997 924 58 × 108 m s-1 is the speed of light, h = 6.626 0693 × 10-34 J s is the Planck constant, and k = 1.380 6505 × 10-23 J K-1 is the Boltzmann constant. The spectral radiance is the energy radiated per unit wavelength by an infinitesimal patch of the black body into an infinitesimal solid angle of space. Hence, its unit is W sr-1 m-2 nm-1.
Figure 2.1 displays the spectral radiance for different temperatures T. It can be seen that black bodies at 300 K radiate primarily in the middle and far IR range. This is the radiation range that is perceived as heat. Therefore, this range of wavelengths is also called thermal IR. The radiation of an object at 1000 K just starts to enter the visible range. This is the red glow that can be seen first when objects are heated. For T = 3000 K, the spectrum is that of an incandescent lamp (see Section 2.1.2). Note that it has a strong red component. The spectrum for T = 6500 K is used to represent average daylight. It defines the spectral composition of white light. The spectrum for T = 10000 K produces light with a strong blue component.
Figure 2.1 Spectral radiance emitted by black bodies of different temperatures. The vertical lines denote the visible range of the spectrum.
Because of the correspondence of the spectra with the temperature of the black body, the spectra also define so-called correlated color temperatures (CIE 15:2004).
2.1.2 Types of Light Sources
Before we take a look at how to use light in machine vision, we will discuss the types of light sources that are commonly used in machine vision.
Incandescent lamps create light by sending an electrical current through a thin filament, typically made of tungsten. The current heats the filament and causes it to emit thermal radiation. The heat in the filament is so high that the radiation is in the visible range of the electromagnetic spectrum. The filament is contained in a glass envelope that contains either a vacuum or a halogen gas, such as iodine or bromine, which prevents oxidation of the filament. Filling the envelope with a halogen gas has the advantage that the lifetime of the lamp is increased significantly compared to using a vacuum. The advantage of incandescent lamps is that they are relatively bright and create a continuous spectrum with a correlated color temperature of 3000-3400 K. Furthermore, they can be operated with low voltage. One of their disadvantages is that they produce a large amount of heat: only about 5% of the power is converted to light; the rest is emitted as heat. Other disadvantages are short lifetimes and the inability to use them as flashes. Furthermore, they age quickly, i.e., their brightness decreases significantly over time.
Xenon lamps consist of a sealed glass envelope filled with xenon gas, which is ionized by electricity, producing a very bright white light with a correlated color temperature of 5500-12 000 K. They are commonly divided into continuous-output short- and long-arc lamps as well as flash lamps. Xenon lamps can produce extremely bright flashes at a rate of more than 200 flashes per second. Each flash can be extremely short, e.g., 1-20 µs for short-arc lamps. One of their disadvantages is that they require a sophisticated and expensive power supply. Furthermore, they exhibit aging after several million flashes.
Like xenon lamps, fluorescent lamps are gas-discharge lamps that use electricity to excite mercury vapor in a noble gas, e.g., argon or neon, causing UV radiation to be emitted. This UV radiation causes a phosphor salt coated onto the inside of the tube that contains the gas to fluoresce, producing visible light. Different coatings can be chosen, resulting in different spectral distributions of the visible light with correlated color temperatures of 3000-6000 K. Fluorescent lamps are driven by alternating current. This results in a flickering of the lamp with the same frequency as the current. For machine vision, high-frequency alternating currents of 22 kHz or more must be used to avoid spurious brightness changes in the images. The main advantages of fluorescent lamps are that they are inexpensive and can illuminate large areas. Some of their disadvantages are a short lifetime, rapid aging, and an uneven spectral distribution with sharp peaks for certain frequencies. Furthermore, they cannot be used as flashes.
A light-emitting diode (LED) is a semiconductor device that produces narrowspectrum (i.e., quasi-monochromatic) light through electroluminescence: the diode emits light in response to an electric current that passes through it. The color of the emitted light depends on the composition and condition of the semiconductor material used. The possible range of colors comprises IR, visible, and near UV radiation. White LEDs can also be produced: they internally emit blue light, which is converted to white light by a coating with a yellow phosphor on the semiconductor. One advantage of LEDs is their longevity: lifetimes larger than 100 000 hours are not uncommon. Furthermore, they can be used as flashes with fast reaction times and almost no aging. Since they use direct current, their brightness can be controlled easily. In addition, they use comparatively little power and produce little heat. The main disadvantage of LEDs is that their performance depends on the ambient temperature of the environment in which they operate. The higher the ambient temperature, the lower the performance of the LED and the shorter its lifetime. However, since LEDs have so many practical advantages, they are currently the primary illumination technology used in machine vision applications.
2.1.3 Interaction of Light and Matter
Light can interact with objects in various ways, as shown in Figure 2.2.
Reflection occurs at the interfaces between different media. The microstructure of the object (essentially the roughness of its surface) determines how much of...
Systemvoraussetzungen
Dateiformat: ePUB
Kopierschutz: Adobe-DRM (Digital Rights Management)
Systemvoraussetzungen:
- Computer (Windows; MacOS X; Linux): Installieren Sie bereits vor dem Download die kostenlose Software Adobe Digital Editions (siehe E-Book Hilfe).
- Tablet/Smartphone (Android; iOS): Installieren Sie bereits vor dem Download die kostenlose App Adobe Digital Editions oder die App PocketBook (siehe E-Book Hilfe).
- E-Book-Reader: Bookeen, Kobo, Pocketbook, Sony, Tolino u.v.a.m. (nicht Kindle)
Das Dateiformat ePUB ist sehr gut für Romane und Sachbücher geeignet – also für „fließenden” Text ohne komplexes Layout. Bei E-Readern oder Smartphones passt sich der Zeilen- und Seitenumbruch automatisch den kleinen Displays an.
Mit Adobe-DRM wird hier ein „harter” Kopierschutz verwendet. Wenn die notwendigen Voraussetzungen nicht vorliegen, können Sie das E-Book leider nicht öffnen. Daher müssen Sie bereits vor dem Download Ihre Lese-Hardware vorbereiten.
Bitte beachten Sie: Wir empfehlen Ihnen unbedingt nach Installation der Lese-Software diese mit Ihrer persönlichen Adobe-ID zu autorisieren!
Weitere Informationen finden Sie in unserer E-Book Hilfe.