Browse Prior Art Database

Chain Codes Below Pixel Resolution

IP.com Disclosure Number: IPCOM000043741D
Original Publication Date: 1984-Sep-01
Included in the Prior Art Database: 2005-Feb-05
Document File: 3 page(s) / 55K

Publishing Venue

IBM

Related People

Koepp, P: AUTHOR

Abstract

Chain codes constitute a means for defining the edge shape of an object standing out against a contrasting background (as is known from numerous publications). The known methods of generating chain codes use an adaptable or a fixed threshold for edge scanning. The result (the chain code) defines the edge shape. (Image Omitted) As the pixels (picture elements) are resolved in four different directions (8 directions being also conceivable, since each pixel has 8 adjacent pixels), the resultant chain reflects the actual edge shape only approximately. This approximation relative to the size of the object is the better; the finer the resolution of an object is represented in the image store. The absolute approximation, however, does not exceed half the pixel length. Thus, the maximum error of the chain code is 0.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 55% of the total text.

Page 1 of 3

Chain Codes Below Pixel Resolution

Chain codes constitute a means for defining the edge shape of an object standing out against a contrasting background (as is known from numerous publications). The known methods of generating chain codes use an adaptable or a fixed threshold for edge scanning. The result (the chain code) defines the edge shape.

(Image Omitted)

As the pixels (picture elements) are resolved in four different directions (8 directions being also conceivable, since each pixel has 8 adjacent pixels), the resultant chain reflects the actual edge shape only approximately. This approximation relative to the size of the object is the better; the finer the resolution of an object is represented in the image store. The absolute approximation, however, does not exceed half the pixel length. Thus, the maximum error of the chain code is 0.5 mm at a resolution of 1 mm pixel length. With respect to image inspection devices, there is the tendency of choosing a large field for the camera, as this approach makes the below-listed advantages obvious: 1. For systematically scanning a large area, fewer images have to be processed, which reduces the processing

time required. 2. The field contains figures in their

entirety, as there are no X-Y table movements and

their inherent inaccuracies. The disadvantage of a large field is, however, that with an image store of, say, 512 x 512, the resolution becomes inaccurate. The loss of accuracy resulting from the coarse resolution may be partly compensated for by interpolating the existing pixel coordinate system to comprise further subcoordinates (subpels) and their grey values. This leads to the generation of a chain code with a higher resolution.

(Image Omitted)

Each coordinate pair of the standard image coordinate system is divided into 16 subcoordinate pairs, which corresponds to a 4 times higher resolution. In such a case, the left upper edge of...