Other Free Encyclopedias » Online Encyclopedia » Encyclopedia - Featured Articles » Contributed Topics from A-E

Edge Detection

image images edges information

Definition: Edge detection is a process of transforming an input digital image into an edge map which can be viewed as a line drawing image.

More specifically, edge detection is a process of transforming an input digital image, color or otherwise, into an edge map which can be viewed as a line drawing image with a spatial resolution identical to that of the input (Figure 1) . An intensity edge map corresponds to sudden changes in brightness which tends to occur at object boundaries. The locations of these changes reflect the shapes of objects. The most popular edge operators generate the edge maps by processing information contained in a local image neighborhood as it is determined by an element of support. These operators i) do not use any prior information about the image structure, ii) are image content agnostic, and iii) are localized, in the sense that the detector output is solely determined by the features obtained through the element of support .

The development of an efficient edge detector, which properly detects the objects’ edges, is a rather demanding task. Assuming scalar signals, such as gray-scale images, edges are commonly defined as physical, photometric and geometrical discontinuities in the image function, as physical edges often coincide with places of strong illumination and reflection changes. Since very often the image contrast is proportional to the scene illumination, identified edges are located in that part of the image where rapid image intensity changes take place. It is therefore reasonable to conclude that edges are the boundaries of distinct, in intensity, image regions. However, the definition of an edge in multichannel images is more complex. Compared to gray-scale images, multichannel images carry additional information contained in the various spectral channels. This information can be used to identify and detect edges in multichannel images such as the conventional color images (three channels) or cDNA microarray images (two channels) . In this case, the boundary between two surfaces with different properties can be determined in more than one way.

Edge-Sensing Mechanism [next] [back] Edelman, Gerald (Maurice)

User Comments

Your email address will be altered so spam harvesting bots can't read it easily.
Hide my email completely instead?

Cancel or