| 摘要 |
The method is characterised in that the method comprises the steps that a computer or several interconnected computers are caused to a) store, in the form of a pixel set in which set each pixel is associated with image information in at least one channel for light intensity, a first image to be classified onto a digital storage medium; b) carry out a first classification of the image, which classification is caused to be based upon the image information of each respective pixel and which classification is caused to associate each pixel with a certain class in a first set of classes, and to store these associations in a first database; c) calculate, for each pixel and for several classes in the first set of classes, the smallest distance in the image between the pixel in question and the closest pixel which is associated with the class in question in the database, and to store an association between each pixel and the calculated smallest distance for the pixel in a second database for each class for which a distance has been calculated; d) carry out a second classification of the data in the second database, which classification is caused to be based upon the smallest distance for each pixel to each respective class, and to associate each pixel to a certain class in a second set of classes; and e) store the classified image in the form of a set of pixels onto a digital storage medium, where each pixel comprises data regarding the association of the pixel to the certain class in the second set of classes, and where the classified image has the same dimensions as the first image.
|
