Digital Image Processing

Course overview

The aim of this course is to provide the attendee with an overview of the fundamentals of image processing and computer vision software techniques as a precursor to further research, study and development within this domain. The course is themed around the development of image processing from a software development perspective and includes practical lab sessions on image processing software development using a C/C++ software environment in addition to lecture/tutorial type sessions.

The most powerful method of sensing available to humans is vision. In computing visual information is represented as a digital image. In order to process visual information in computer systems we need to know about processing digital images. Here we focus upon the task of low-level visual processing in the digital form and how to implement such techniques in software.

The course will be delivered as a series of short lectures, tutorials and practical lab sessions in which students will develop a range of practical image processing algorithms using the C/C++ software environment. Presented lecture content is supported by "live" in-lecture image processing demos, the program source code for which is made available to delegates to form part of the practical lab sessions. Lab sessions are PC based using C/C++ programming to perform image processing from images, videos and PC connected cameras.

Objectives

On successful completion of this module, the attendee will be able to:

• understand digital image representations
• understand and implement a range of image transforms
• understand and implement image processing in the frequency domain
• implement basic feature extraction and matching
• understand the effects of noise on all aspects of digital imaging and implement a range of noise reduction filtering approaches
• understand and appreciate the broader application implications of a given image processing solution for a particular industrial application.

Topics

Programming (in C/C++ with Open CV)

• image loading and display
• applying image transforms
• image manipulation
• writing images and videos
• live image processing from a connected camera

Imaging: applications/representation/hardware/sampling/noise

Image geometry and locality

Operations upon images

Mathematical Background: Camera projection/convolution

Transformation

• arithmetic/logical operations
• thresholding/Fourier Transform
• image convolution/correlation matching/de-convolution
• using image histograms for comparison
• high pass filtering/low pass filtering/band-pass filtering
• colour transforms - RGB and HSV colour spaces

Enhancement

• logarithmic and exponential transforms/gamma correction
• histogram transforms: contrast stretch/equalisation/matching
• homomorphic filtering
• edge enhancement
• noise characteristics and noise reduction filtering

Who should attend?

The course will be of interest to those in the following job roles:

• imaging engineer
• vision processing engineer
• software engineer
• programmer
• vision or image processing scientist
• medical imaging professional
• embedded engineer
• embedded software engineer.

A working knowledge of C/C++ programming is assumed for course attendees.

Previous delegates have come from various industries including:

• industrial inspection/manufacturing
• security/defence/transport imaging
• robotics
• weather and environmental monitoring (remote sensing)
• medical imaging.