Image analysis

This paper provides an overview of modern image analysis techniques pertinent to materials science. The usual approach in image analysis contains two basic steps: first, the image is segmented into its constituent components (e.g. individual grains), and second, measurement and quantitative analysis are performed. Usually, the segmentation part of the process is the harder of the two. Consequently, much of the paper concentrates on this aspect, reviewing both fundamental segmentation tools (commonly found in commercial image analysis packages) and more advanced segmentation tools. There is also a review of the most widely used quantitative analysis methods for measuring the size, shape and spatial arrangements of objects. Many of the segmentation and analysis methods are demonstrated using complex real-world examples. Finally, there is a discussion of hardware and software issues. 42 refs., 17 figs