Lecture 1: Introduction to robot vision
In this lecture we discuss the problem of sensing for robots, absolute and relative, and the particular advantages of vision for animals and robots. We then have a brief introduction to using MATLAB and the Machine Vision Toolbox for some simple image processing applications.
Lecture 2: Image processing
Introduction to digital images (greyscale), image processing, histograms, thresholds, smoothing, moments, blobs, area and centroid.
To get the Matlab toolbox used here, visit petercorke.com.
Lecture 3: blob features
Connectivity analysis, position and area of multiple blobs, topological relationship between blobs.
Lecture 5: Color
In this lecture we discuss colour, why its important, and how data in a color image is organized in planes, how 3 channel color can be considered as intensity plus a 2-dimensional chromaticity component, and we use chromaticity to classify pixels in an image.
(Note lecture 4 didn’t happen)
Lecture 6: The source of colour
In this lecture we look at the phenomena underlying the human perception of color, the spectral characteristics of light entering the eye which is a function of the light source and the object it reflects from. We discuss blackbody radiators, Planck’s law and several non-Planckian sources such as LEDs, lasers and compact fluorescent lights. We also discuss luminosity which is the ratio of light source power to perceived brightness.
Lecture 7: blob features and mathematical morphology
In this lecture we go back to considering the features of a blob, previously we discussed how to compute its position and area, here we consider ways to describe shape and orientation. We then discuss mathematical morphology, a powerful image processing technique that filters blobs according to their size and shape.
Lecture 8: image formation
In this lecture we look at how images are formed, from light reflecting off points in the world and using a pinhole or lens to an image being created on an image plane. The pinhole has very simple geometry but in practice leads to very dim images so we use a lens to gather more light, but this leads to problems with focus. Both can be modeled using a perspective projection which maps 3 dimensions down to 2, and the consequence is that lines in the world are not necessarily parallel in the image, and circles may appear as ellipses. Finally we touch again on the problem of focus and introduce the concept of light field cameras.
Lecture 9: Image geometry and planar homography
In this lecture we discuss in more detail the equation of image formation, particularly their expression in matrix form using homogeneous coordinates. We then introduce the planar homography, a mapping from points in a world plane to the image plane, and show some examples of how this can be used for real problems.
Lecture 10: perceiving depth & 3D reconstruction
In this last lecture we look at the many mechanisms we use to perceive the distance of objects in the world. We then look at stereo vision and the horizontal shift, disparity, in world points when we translate a camera and which is a function of depth.
(Source: Peter Corke)