Image analysis: an overview of developments

Performing image analysis involves digitizing and analyzing an image to characterize the features in it. A few researchers use image analysis to solve specific problems by writing their own computer programs, but widespread usage requires the availability of commercial image analyzers. Manufacturers have built a wide variety of image analyzers whose price and capabilities vary by a large factor. All image analyzers consist of image input, image digitizing, image processing, and data output. The image input can be (i) a digitizer tablet on which an image is drawn with a special stylus; (ii) a TV camera for collecting an image from an optical microscope or an epidiascope; (iii) a SEM interface for collecting a SEM image; and (iv) an image from computer memory. Image processing varies from simple processing such as determining the area % of a specific phase, to measuring a wide variety of parameters for each feature, as well as determining the reflectance and transmittance of features observed with an optical microscope. Furthermore, the electron beam of a SEM can be steered to scan each feature and to identify the feature on the basis of its composition. Image analyzers, therefore, vary from the simplest one which uses a digitizer tablet and can perform only a few measurements such as area % and size analysis of the features, to a fully equipped one that can automatically identify all the phases in the image and characterize each feature by using a wide variety of parameters.