Chapter 3

Fusion in Image Processing

In the same way as the previous chapter described the specificities of fusion techniques applied to signal processing, this chapter will focus on the specificities of fusion in image processing. We will go back to the general definitions provided in Chapter 1 and discuss them in this particular context. We wish to emphasize the specific nature of images and their representation in fusion problems, and insist on what makes fusion in image processing different from most of the other application fields in fusion.

3.1. Objectives of fusion in image processing

Images appeared of course very early on as important sources of information for existing information fusion systems and data fusion systems have used images. Let us consider, for example, a comprehensive tracking application for ecological situations. It requires remote sensing to provide weather information. The data provided by the image can then be integrated in a physical model by estimating, for each pixel, the cloud cover. We can include in a thermodynamics balance equation the level of water vapor estimated this way for each point inside the image.

However, this is not the type of application that has led to the original field of image fusion. We should look instead at the practice of image interpretation experts, where we will find the models that image processors have tried to copy, from widely different areas of society. Here are two examples, but we could easily find many other and ...

Get Information Fusion in Signal and Image Processing: Major Probabilistic and Non-Probabilistic Numerical Approaches now with the O’Reilly learning platform.

O’Reilly members experience live online training, plus books, videos, and digital content from nearly 200 publishers.