Scribd
Upload
20 views3 pages

Basic building blocks of Machine vision

Machine vision is essential in manufacturing for quality inspections and product traceability, involving tasks like object identification, position detection, completeness checking, shape inspection, and surface inspection. The document outlines the components of a typical machine vision system, including cameras, illumination, and software, and emphasizes the multidisciplinary expertise required for development. The book aims to provide a solid foundation in hardware components and algorithms relevant to machine vision applications.

Uploaded by

ddeivasigamani E&I-HICET
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
20 views3 pages

Basic building blocks of Machine vision

Machine vision is essential in manufacturing for quality inspections and product traceability, involving tasks like object identification, position detection, completeness checking, shape inspection, and surface inspection. The document outlines the components of a typical machine vision system, including cameras, illumination, and software, and emphasizes the multidisciplinary expertise required for development. The book aims to provide a solid foundation in hardware components and algorithms relevant to machine vision applications.

Uploaded by

ddeivasigamani E&I-HICET
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

1

1
Introduction

Machine vision is one of the key technologies in manufacturing because of increas-


ing demands on the documentation of quality and the traceability of products. It is
concerned with engineering systems, such as machines or production lines, that can
perform quality inspections in order to remove defective products from production
or that control machines in other ways, e.g., by guiding a robot during the assembly
of a product.
Some of the common tasks that must be solved in machine vision systems are as
follows (Fraunhofer Allianz Vision, 2003):
• Object identification is used to discern different kinds of objects, e.g., to control
the flow of material or to decide which inspections to perform. This can be based
on special identification symbols, e.g., character strings or bar codes, or on specific
characteristics of the objects themselves, such as their shape.
• Position detection is used, for example, to control a robot that assembles a product
by mounting the components of the product at the correct positions, such as in a
pick-and-place machine that places electronic components onto a printed circuit
board (PCB). Position detection can be performed in two or three dimensions,
depending on the requirements of the application.
• Completeness checking is typically performed after a certain stage of the assembly
of a product has been completed, e.g., after the components have been placed onto
a PCB, to ensure that the product has been assembled correctly, i.e., that the right
components are in the right place.
• Shape and dimensional inspection is used to check the geometric parameters of a
product to ensure that they lie within the required tolerances. This can be used
during the production process but also after a product has been in use for some
time to ensure that the product still meets the requirements despite wear and tear.
• Surface inspection is used to check the surface of a finished product for imperfec-
tions such as scratches, indentations, protrusions, etc.
Figure 1.1 displays an example of a typical machine vision system. The object (1)
is transported mechanically, e.g., on a conveyor belt. In machine vision applications,
we would often like to image the object in a defined position. This requires mechani-
cal handling of the object and often also a trigger that triggers the image acquisition,
e.g., a photoelectric sensor (4). The object is illuminated by a suitably chosen or

Machine Vision Algorithms and Applications, Second Edition. Carsten Steger, Markus Ulrich, and Christian Wiedemann.
© 2018 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2018 by Wiley-VCH Verlag GmbH & Co. KGaA.
2

Hole Radius OK
2 1 3.01 mm ✔
2 2.99 mm ✔

...
3 3.02 mm ✔

.
8 4 2.87 mm ✘
Overall ✘
7 9

3 3

6 10
1 4 5

11

13
12

Figure 1.1 The components of a typical machine vision system. An image of the object to
be inspected (1) is acquired by a camera (2). The object is illuminated by the illumination
(3). A photoelectric sensor (4) triggers the image acquisition. A computer (5) acquires the
image through a camera–computer interface (6), in this case a frame grabber. The
photoelectric sensor is connected to the frame grabber. The frame grabber triggers the
strobe illumination. A device driver assembles the image (7) in the memory of the
computer. The machine vision software (8) inspects the objects and returns an evaluation
of the objects (9). The result of the evaluation is communicated to a PLC (11) via a digital
I/O interface (10). The PLC controls an actuator (13) through a fieldbus interface (12).
The actuator, e.g., an electric motor, moves a diverter that is used to remove defective
objects from the production line.

specially designed illumination (3). Often, screens (not shown) are used to prevent
ambient light from falling onto the object and thereby lowering the image quality.
The object is imaged with a camera (2) that uses a lens that has been suitably se-
lected or specially designed for the application. The camera delivers the image to a
computer (5) through a camera–computer interface (6), e.g., a frame grabber. The de-
vice driver of the camera–computer interface assembles the image (7) in the memory
of the computer. If the image is acquired through a frame grabber, the illumination
may be controlled by the frame grabber, e.g., through strobe signals. If the camera–
computer interface is not a frame grabber but a standard interface, such as IEEE 1394,
USB, or Ethernet, the trigger will typically be connected to the camera and illumina-
tion directly or through a programmable logic controller (PLC). The computer can
be a standard industrial PC or a specially designed computer that is directly built into
the camera. The latter configuration is often called a smart camera. The computer
may use a standard processor, a digital signal processor (DSP), a field-programmable
gate array (FPGA), or a combination of the above. The machine vision software (8)
inspects the objects and returns an evaluation of the objects (9). The result of the
evaluation is communicated to a controller (11), e.g., a PLC or a distributed control
3

system (DCS). Often, this communication is performed by digital input/output (I/O)


interfaces (10). The PLC, in turn, typically controls an actuator (13) through a com-
munication interface (12), e.g., a fieldbus or serial interface. The actuator, e.g., an
electric motor, then moves a diverter that is used to remove defective objects from
the production line.
As can be seen from the large number of components involved, machine vision
is inherently multidisciplinary. A team that develops a machine vision system will
require expertise in mechanical engineering, electrical engineering, optical engineer-
ing, and software engineering.
To maintain the focus of this book, we have made a conscious decision to focus
on the aspects of a machine vision system that are pertinent to the system until the
relevant information has been extracted from the image. Therefore, we will forgo
a discussion of the communication components of a machine vision system that are
used after the machine vision software has determined its evaluation. For more infor-
mation on these aspects, please consult Caro (2003); Berge (2004); Mahalik (2003).
In this book, we will try to give you a solid background on everything that is re-
quired to extract the relevant information from images in a machine vision system.
We include the information that we wish someone had taught us when we started
working in the field. In particular, we mention several idiosyncrasies of the hard-
ware components that are highly relevant in applications, which we had to learn the
hard way.
The hardware components that are required to obtain high-quality images are de-
scribed in Chapter 2: illumination, lenses, cameras, and camera–computer interfaces.
We hope that, after reading this chapter, you will be able to make informed decisions
about which components and setups to use in your application.
Chapter 3 discusses the most important algorithms that are commonly used in ma-
chine vision applications. It is our goal to provide you with a solid theoretical foun-
dation that will help you in designing and developing a solution for your particular
machine vision task.
To emphasize the engineering aspect of machine vision, Chapter 4 contains a
wealth of examples and exercises that show how the machine vision algorithms dis-
cussed in Chapter 3 can be combined in non-trivial ways to solve typical machine
vision applications.

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy