Edge detection
The edge detection (english edge detection), rare edge extraction, is part of a segmentation of elements in the image processing. They tried to separate flat areas in a digital image of each other, if they differ along straight or curved lines sufficiently in color or gray scale, brightness or texture. Special edge operators to detect the transitions between these areas and mark them as edges. At the same time, a single, homogeneous region is to be detected as such and are not shared by a good edge detector by an edge in the two surfaces.
An edge detector is calculated from an image a corresponding edge image in which all edges are seen accordingly. To achieve this, each pixel is updated by a calculation by means of a matrix. However, this is not a matrix in the mathematical sense, but a filter mask. So you does not use matrix operations, but the discrete convolution.
The main difference between the various edge filters is that they use different filter masks.
Operators
The most common edge filter (also edge operators ) are:
- Sobel operator
- Scratching operator
- Laplace filter (see also Laplace operator )
- Prewitt operator
- Roberts operator
- Kirsch operator
- Canny algorithm
- Contrast Enhancer
- Active contour ( Snake )
A major problem of the edge detection, particularly three-dimensional objects, is the difference of reflection edges, based on characteristics of the object, lighting and edges that are based on characteristics of the light ( such as shadows, light cone).
For humans, edge detection is an essential requirement for visual object recognition. His visual system performs this task mainly with lateral inhibition. Thus even the finest line structures are recognizable from a distance, such as thin wires or remote Wegverläufe. However, optical illusions are possible, as some in the so-called Martian canals or the Mach stripes.
Many graphics programs already have a filter for the detection of contour edges ( engl. edge detection). This will, inter alia, in the photogrammetry and cartography used to detect from aerial photographs exact object or terrain edges. For digital images, the gray value of each picture element ( pixel) compared with that of the neighboring elements or differentiated according to the two image coordinates. The contact edges of the second discharge line ( variation of the brightness gradient ) usually most apparent.
- General Psychology
- Edge detection