Action observed and controlled vision system feedback loop
2 |
|
1 Introduction |
---|---|---|
1.1 |
It is necessary to assume at the outset that no image acquisition or display is needed. Is there then any difference in the architecture of a computer vision system from a general computer? Yes? No? The answer is no, as a general-purpose computer is built only to be used to compute a wide range of tasks. It is thus only a question of how effi-ciently modern personal computer (PC) and workstation architectures can be used for computer vision tasks. The answer is yes, as for many applications it is required to put a stream of image data at a specified rate through the computer system. This requires so-called real-time computer hard- and software. Nowadays three trends are observable with respect to the platforms for computer vision applications.
Although a PC with a frame grabber and a camera can be regarded as a simple computer vision system, it generally is more than that. This can be seen by comparing computer vision systems with human or other biological vision systems. While it makes not much sense to imitate a biological system with a technical system, it is very useful to compare them on the basis of function modules as shown in Table 1.1.
1.1 Computer vision architecture 3
Visualization | Passive, mainly by re- | Passive and active (controlled il- | |||
flection | of | light | from | lumination) using electromagnetic, |
opaque surfaces
particulate, and acoustic radiation (Volume 1, Chapters 3 and 6)
optical | sys- | |||
---|---|---|---|---|
|
|
|||
Motorized apertures, filter wheels, | ||||
|
||||
|
Autofocus systems based on vari- |
|
||||||
---|---|---|---|---|---|---|
|
8- | and | 16-bits; | logarith- |
|
---|
eras (Chapter 9)
and analysis |
|
|
|