Author: Paul Mather & Magaly Koch
Publisher: Wiley-Blackwell, 2010
Aimed at: Specialists in biology, geography, meteorology, etc
Pros: Focused on a particular use of image processing
Cons: Doesn't cover mathematical techniques in depth
Reviewed by: Mike James
In this day of GIS, Google Earth and Maps we could all do with a quick course in remote sensing.
Remotely sensed imagery is an odd subject that requires knowledge from a range of different disciplines. You need to be a physicist to follow the way the images are captured. You need to be a mathematician to understand how they can be processed. You need to be a computer expert to implement the processing and finally you need to be an expert in one of the subjects that are the focus of the imaging - biology, geography, meteorology and so on.
This book is mostly aimed at the subject specialist wanting to find out about remote sensed images. It could also be of use to the programmer wanting to know what sorts of things are standard algorithms and what might yet need to be invented.
The book starts off from the basic principles of physics involved in imaging. Then on to the specifics of the various platforms used to capture images - Spot, Landsat, etc.. After this the book moves on to what you might call classical image processing. Chapter 5 is on enhancement and covers histogram-based contrast enhancement etc. The next chapter deals with image transformations - from addition of images, though PCA, Fourier transform, wavelets, change detection and image fusion. If you have a background in image processing then you might find the treatment a little shallow but its a good summary and the final sections on change detection and fusion should be new to you. Chapter 7 deals with classical filtering, chapter 8 with classification including k means, statistical classification and neural networks. Chapter 9 deals with advanced processing mostly based on physics - interferometry, spectroscopy and Lidar. The final chapter puts everything into the context of Geographical Information Systems GIS.
Many of the examples in the book make use of the MIPS software written by one of the authors and available for download from the book's website. There are also some sample images that you can work with.
The book is well written and very nicely produced. It forms a good introduction to image processing as applied to satellite photos and as such will be of help if you are trying to do a range of GIS tasks. If the book has a flaw it is that it fails to tackle some of the more advanced mathematical techniques but some readers might find this a plus point.