Short Course on Computer Vision for Automated
Industrial Inspection
A five day course to be held at the University of Surrey, Guildford.
16th-20th September 1996.
This course has been organised by Dr M Petrou and Professor J Kittler, both
of whom have extensive experience on many aspects of industrial inspection.
All the course lecturers have been directly involved in automatic
inspection projects and they are all members of the Vision, Speech and
Signal Processing Group, one of the largest groups of its kind in the
country.
Course Objectives
- To give the necessary background for engineers to be able to
appreciate and judge for themselves proposed solutions and approaches to their
industrial inspection problems.
- To familiarise engineers with the latest theoretical
and technical developments in the area through hands on experience.
- To give an idea of the benefits and limitatons of
current technology within the context of automatic
industrial inspection.
- To give an overview of relevant hardware platforms
and software environments that are commercially available.
Provisional Program
Day One:
Introduction
- Industrial inspection problems (surface properties, shape)
- Benefits of automation. What is possible. Basic system components.
- Overview of EU programmes in this area. Course Outline.
Basic Image Processing I
- Image Transforms
(Fourier, Walsh, Hadamard)
Sensors
- Basic principles of image acquisition
(digitisation and quantisation).
- Basic principles of sensors available. The CCD
camera. Resolution, Linearity, Sensitivity.
Corrections and calibration. What is currently
available. Frame grabbers.
Practical
Day Two:
Basic Image Processing II
- Image enhancement and restoration.
- Filtering, Thresholding, Edge Detection
Texture: An overview
- What is texture; ways of quantifying it. Regular
versus stochastic textures. Co-occurrence
matrices, Laws' masks. Basic concepts behind
more sophisticated representations (Gabor,
Wavelets, Wigner).
Texture defect detection
- Defect detection in regular textures (alignment
and registration). Crack detection. Blob
detection. Case studies.
Practical
Day Three:
Decision Making I
- Statistical models. Hypothesis testing.
Classification of defects. Feature selection.
Clustering.
Colour
- Colour perception, colour spaces, illumination
correction. Colour and texture
Colour grading
- Similarity measures. Perceptual corrections.
Practical
Day Four:
Decision Making II
- Neural networks. Estimations.
Performance assessment.
2D Shape inspection
- 2D global shape description.
Moments, Fourier Descriptors, Hough Transform.
Local Feature Focus.
Software packages and hardware platforms
- Currently available software packages.
- Hardware platforms (PC and non PC-based).
Dedicated pipeline processors.
Properties and limitations.
Practical
Day Five:
Range sensors
- Physical principles behind the various
techniques. Commercially available
technologies. Comparisons.
Range data algorithms
- From range data to CAD models: reverse
engineering CAD-based inspection.
Shape-based inspection and control.
Practical
Cost
The price per person which includes coffee, lunch, tea and printed
course notes is
- £975 standard rate.
- £925 before 16 August 1996
As an educational institution we do not charge VAT.
Contact Person
Patricia Yehia, Phone: 01483 25 9837, Fax 01483 34139
Meet the Staff!
Useful Information
Other links
Small Print
The University reserves the right to cancel any course on the grounds of
insufficient numbers or other reasons outside our control. In the event of
cancellation by the University, a full refund will be given.
In the event of cancellation by a delegate, a refund of 75% will be given
provided written notice is received two weeks prior to the event. If less than
this notice is received, a refund of 50% will be given.
M.Mirmehdi@ee.surrey.ac.uk
June 1996