Image processing

Image processing is a general term for the manipulation of a digitized image to produce an enhanced or more con-venient version. Some of the earliest applications were in the military (aerial and, later, satellite reconnaissance) and in the space program. The military and space programs had a great need for extracting as much useful information as possible from images that were often gathered under extreme or marginal conditions. They also needed to make cameras and other hardware components simultaneously more compact and more efficient, and generally had the funds to pay for such specialized developments.

Once developed, higher-quality image processing sys-tems found their way into other applications such as domes-tic surveillance and medical imaging. The development of cameras that could directly turn light into digitized images (see photography, digital) made image processing seam-less by avoiding the necessity of scanning images from tra-ditional film.

Image processing applications can be divided into three general categories: enhancement, interpretation, and maintenance.

Enchancement

Enhancement includes bringing out objects of interest (such as enemy vehicles or a particular rock formation on Mars) from the surrounding background by enhancing contrast or applying appropriate filters to block out the background. More sophisticated filters can also be used to compensate for defects in the original image, such as “red-eye,” blur, and loss of focus. Today’s image processing programs, such as the popular Adobe Photoshop, make relatively sophisti-cated image manipulation techniques available to interested amateurs as well as professionals. More sophisticated image enhancement techniques include the creation of 3D images based upon the differences calculated from a number of photos shot from slightly different angles.

A considerable amount of image enhancement takes place even before the photo is taken. Today’s versatile cam-eras (see photography, digital) include a variety of modes that are preset for different scenarios such as indoor portrait or low light. After the picture is taken, photo management programs (often bundled with the camera or even included in the operating system) not only help organize photos, but also provide simple ways to crop or enhance them.

Interpretation

Interpretation refers to manipulation designed to help human observers obtain more and better information from the image. For example, “false color” can be used to heighten otherwise imperceptible color differences in the original image, or to translate nonvisual information (such as heat or radio emission levels) into visual terms. Artificial intelligence algorithms can also be employed to automatically analyze images for features of interest (see pattern recognition and computer vision). In fields such as military reconnaissance this might allow a high volume of imagery to be prescreened, with images meet-ing certain criteria “flagged” for the attention of human interpreters.

Maintenance

Maintenance includes archiving of images, often with the aid of compression to reduce the amount of storage space required (see data compression). It can also include the restoration of images that may have been degraded (as from chemical decomposition of stored film.) This can be done either by creating a reversible mathematical model of the degradation process (thus, for example, restoring colors that have changed through oxidation or other processes) or by creating a model of how the image was formed in the first place and comparing its output to the existing image.