Means and method of processing optical image edge data
US4553260A · kind A · utility
Assignee
Inventors
Key dates
| Filing date | Mar 18, 1983 |
| Grant date | Nov 12, 1985 |
| Priority date | — |
| Expiry date | Mar 18, 2003 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06T7/12
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
Means and methods are provided for processing the quantity Si which corresponds to a point i in space where Si=(Xi.sup.2 +Yi.sup.2).sup.1/2 and Xi and Yi are quantities lying along mutually orthogonal axes in the space. Si is approximated by the expression A.vertline.Xi.vertline.+B.vertline.Yi.vertline. where A and B are constant coefficients selected to minimize the error in value of Si according to a selected error computation. A and B can be approximated as a binary power series with each series including a preselected number of terms. Another embodiment does not require means for approximating A and B as binary power series, but segments the coordinate space into m sectors and computes and Am and Bm for each of said m sectors. A particularly advantageous application of the invention is its use in a device and method for processing edge data concerning an edge in an optical image where Si is the Sobel, Prewitt or Roberts square root edge operators.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.