Decoding distorted symbols

• US 20110101101A1
• Filed: 01/06/2011
• Published: 05/05/2011
• Est. Priority Date: 12/20/2005
• Status: Active Grant

• Current Assignee: Cognex Corporation
• Original Assignee: Cognex Technology & Investment Corporation (Cognex Corporation)
• Inventors: Ye, Xiangyun, Nadabar, Sateesha
• Granted Patent: US 8,322,620 B2
• Time in Patent Office: Days
• US Class Current: 235/462.11
• CPC Class Codes: G06K 7/14 using light without selection of wavelength, e.g. sensing reflected white light G06K7/10831 - G06K7/1097 take precedence
• CPC Class Codes: G06T 3/00 Geometric image transformations in the plane of the image

Abstract

A method is provided for reading distorted optical symbols using known locating and decoding methods, without requiring a separate and elaborate camera calibration procedure, without excessive computational complexity, and without compromised burst noise handling. The invention exploits a distortion-tolerant method for locating and decoding 2D code symbols to provide a correspondence between a set of points in an acquired image and a set of points in the symbol. A coordinate transformation is then constructed using the correspondence, and run-time images are corrected using the coordinate transformation. Each corrected run-time image provides a distortion-free representation of a symbol that can be read by traditional code readers that normally cannot read distorted symbols. The method can handle both optical distortion and printing distortion. The method is applicable to “portable” readers when an incident angle with the surface is maintained, the reader being disposed at any distance from the surface.

Claims

1-35. -35. (canceled)

36. A method for decoding distorted symbols, the method comprising:

acquiring a run-time image of a symbol, the symbol showing distortion in the run-time image;

obtaining a coordinate transformation that describes the distortion shown in the image of the symbol; and

using a processor, where the processor corrects distortion shown in the run-time image using the coordinate transformation so as to provide a corrected image;

running a run-time symbol reader on the corrected image; and

wherein the coordinate transformation is constructed using a correspondence between at least two 2D coordinates of points within a train-time image and two 2D coordinates of points within a 2D code symbol, the correspondence being provided by a symbol reader that can detect and tolerate distortion of the 2D code symbol.

37. The method of claim 36 wherein the run-time symbol reader includes the processor.