Indications exist that lines, vertices, and other features based on lines, are very important for perception [Mar80]. A typical part of image segmentation is the detection of edges (c.f. Sect. 6.5). The automatic detection of line features in images usually requires several processing stages. Edge detection operators are applied to every pixel in the image. These operators check the local neighborhood for evidence of an edge. They return a measure for the likelihood of an edge at this point of the image as well as a guess of its orientation. The result is called an “edge image” (Sect. 14.6). In Chapter 21 we will further process edge images to obtain lines. This will transform edge images into more abstract geometric objects. This chapter gives a first insight into edge detection methods in gray—level images, partially based on [Brü90].
Unable to display preview. Download preview PDF.