Detecting bar code candidates
AssigneeAccu-Sort Systems, Inc.
InventorsKurt Hecht, Milorad Neskovic, Vasanth Shreesha, Edward Cohen
Life of this patent
Estimating the term…
Abstract
An area of interest is scanned to produce pixels of the scanned area. The pixels are divided into a plurality of blocks. For each block, an orientation of a spanning line or a lack of orientation is determined. Contiguous ones of the blocks are clustered, when their orientation is within a certain tolerance level. Each cluster has a length and a width. The length is longer than the width. If the orientation of a cluster's block is perpendicular to the length of that cluster, that cluster is selected as a bar code candidate.
Claims19
A method for detecting bar code candidates, the method comprising:
- scanning an area of interest to produce pixels of the scanned area;
- dividing the pixels into a plurality of blocks;
- for each block, determining an orientation of a spanning line or a lack of orientation when a spanning line is not present;
- clustering contiguous ones of the blocks having their orientation within a certain tolerance level, each cluster having a length and a width, the length longer than the width;
- and
- selecting candidate bar codes from the clusters, by determining if the orientation of each cluster's blocks is perpendicular to the length of the cluster.
A bar code scanner comprising:
- a scanner for scanning an area of interest to produce pixels of the scanned area;
- a processor for executing instructions including:
- scanning an area of interest to produce pixels of the scanned area;
- dividing the pixels into a plurality of blocks;
- for each block, determining an orientation of a spanning line or a lack of orientation when a spanning line is not present;
- clustering contiguous ones of the blocks having their orientation within a certain tolerance level, each cluster having a length and a width, the length longer than the width;
- and
- selecting candidate bar codes from the clusters, by determining if the orientation of each cluster's blocks is perpendicular to the length of the cluster.
Dependencies are inferred from the claim text.