CHAPTER 2 I NDUSTRIAL V ISION

In an industrial vision camera, sensors and a processor make

decisions on what the camera sees and then relays information
back to the handling system. It introduces automation into the
production process in all levels of product manufacturing in
different kind of industries. An industrial inspection system
computes information from raw images according to the
steps of image acquisition, image processing, feature extraction,
and decision
making. An industrial camera is designed according to high
standards with repeatable performance, and it must be robust to
withstand the demands
of harsh industrial environments. The industrial vision
applications are classified into four types: dimensional quality,
surface quality, structural quality, and operational quality.

A typical industrial vision system.

 This Figure illustrates the structure of a typical industrial vision
system. First, a computer is employed for processing the
acquired images. This is achieved by applying special purpose
image processing analysis and classification software.
Images are usually acquired by one or more cameras placed at
the scene under inspection. The positions of the cameras are
usually fixed. In most cases, industrial automation systems are
designed to inspect only known objects at fixed positions.
The scene is appropriately illuminated and arranged in order to
facilitate the reception of the image features necessary for
processing and classification. These features are also known in
advance. When the process is highly time constrained or
computationally intensive and exceeds the processing
capabilities of the main processor, application specific hardware
(e.g., DSPs, ASICs, or FPGAs) is employed to reduce the
problem of processing speed. The results of this processing can
be used to:
1/ Control a manufacturing process (e.g., for guiding robot arms
placing components on printed circuits, painting surfaces, etc.)

2/ Propagate to other external devices (e.g., through a network

or other type of interface like FireWire) for further processing
(e.g.,classification)

3/ Characterize defects of faulty items and take actions for

reporting and correcting these faults and replacing or removing
defective parts from the production line

4/ The requirements for the design and development of a

successful machine vision system vary depending on the
application domain and are related to the tasks to be
accomplished, environment, speed, and so on. For example, in
machine vision inspection applications, the system must be able
to differentiate between acceptable and unacceptable variations
or defects in products. While in other applications, the system
must enable users to solve the guidance tasks, alignment tasks,
measurement, and also the assembly verification tasks.

The important attributes of an industrial machine vision

inspection system are such as, flexibility, efficiency in
performance, speed and cost, reliability and robustness. In order
to design a system that maintains these attributes it is
important to clearly define its required outputs and the available
inputs. Typically, an industrial inspection system computes
information from raw images according to the following
sequence of steps:
1/ Image acquisition: Images containing the required
information are acquired in digital form through cameras,
digitizers, and so on.

2/Image processing: Once images have been acquired, they are

filtered to remove background noise or unwanted reflections
from the illumination system. Image restoration may also be
applied to improve image quality by correcting geometric
distortions introduced by the acquisition system (e.g., the
camera).
3/Feature extraction: A set of known features, characteristic of
the application domain, is computed, probably with some
consideration for nonoverlapping or uncorrelated features, so
that better classification can be achieved. Examples of such
features include size, position, contour measurement via edge
detection, and linking, as well as and texture measurements on
regions. Such features can be computed and analyzed by
statistical or other computing techniques (e.g., neural networks
or fuzzy systems). The set of computed features forms the
description of the input image.

4/Decision making: Combining the feature variables into a

smaller set of new feature variables reduces the number of
features. While the number of initial features may be large, the
underlying dimensionality of the data, or the intrinsic
dimensionality, may be quite small. The first step in decision
making attempts to reduce the dimensionality of the feature
space to the intrinsic dimensionality of the problem. The
reduced feature set is processed further as to reach a decision.
This decision, as well as the types of features and measurements
(the image descriptions) computed, depends on the application.
For example, in the case of visual inspection during production
the system decides if the produced parts meet some quality
standards by matching a computed description with some known
model of the image (region or object) to be recognized. The
decision (e.g., model matching) may involve processing with
thresholds, statistical, or soft classification.
CHAPTER 3 MEDICAL VISION
Vision technology in medicine allows doctors to see interior
portions of the body for easy diagnosis, making it possible to
perform minimally invasive surgery when surgery is necessary.
Image processing for medical applications include endoscopy,
CT, ultrasonic imaging system, MRI, X-ray imaging, MIS,
PACS, Ophthalmology, ICG imaging, corneal image analyzer,
fundus image analyzer, facial recognition to determine pain,
patient activity detection, peripheral vein imaging, and
stereoscopic microscope. CNN takes an input image of raw
pixels and transforms them via convolution layers, rectified
linear unit layers, and pooling layers. This feeds into fully
connected layers that assign class scores, thus classifying the
input into the affected area.

in medical imaging there are four key problems:

1. Segmentation—automated methods that create patient-pecific
models of relevant anatomy from images;
2. Registration—automated methods that align multiple data sets
with each other;
3. Visualization—the technological environment in which
image-guided procedures can be displayed;
4. Simulation—software that can be used to rehearse and plan
procedures, evaluate access strategies, and simulate planned
treatments.

1/ Advantages of Digital Processing for Medical Applications

􀁑 Digital data will not change when it is reproduced any
number of times
and retains the originality of the data.
􀁑 Digital image data offers a powerful tool to physicians by
easing the search for representative images.
􀁑 Images are displayed immediately after acquisition.
􀁑 Enhancement of images makes them easier for the physician
to interpret.
􀁑 Allows for quantifying changes over time.
􀁑 Provides a set of images for teaching to demonstrate examples
of diseases or features in any image.
􀁑 Allows for quick comparison of images.

2/ Digital Image Processing Requirements for Medical

Applications
􀁑 Interfacing analog outputs of sensors such as microscopes,
endoscopes, ultrasound, and so on to digitizers and in turn to
digital image processing systems
􀁑 Image enhancements
􀁑 Changing density dynamic range of B/W images
􀁑 Color correction in color images
􀁑 Manipulation of colors within an image
􀁑 Contour detection
􀁑 Area calculations of the cells of a biomedical image
􀁑 Display of image line profile
􀁑 Restoration of images
􀁑 Smoothing of images
􀁑 Registration of multiple images
􀁑 Construction of 3-D images from 2-D images
􀁑 Generation of negative images
􀁑 Zooming of images
􀁑 Pseudo coloring
􀁑 Point-to-point measurements
􀁑 Getting relief effect
􀁑 Removal of artifacts from the image

3/Advanced Digital Image Processing Techniques in Medical

Vision
􀁑 Neural network-based image processing
􀁑 Statistical approach for texture analysis
􀁑 Segmentation in color and B/W images
􀁑 Expert system-based image processing.
􀁑 Application of object-oriented programming techniques in
imageprocessing environments
􀁑 Shape in machine vision
􀁑 Multispectral classification techniques
􀁑 Auto-focus techniques for MRI images
􀁑 Threshold technique for finding contours of objects
􀁑 Sequential segmentation technique to detect thin vessels in
medical images and hair-line cracks in nondestructive testing
􀁑 Fractal method for texture classification
􀁑 Data compression techniques using fractals and discrete
cosine transformers
􀁑 Image restoration methods using point-spread functions and
Wiener filter, and so on.