Image Recognition Apparatus and Storage Medium

• US 20160155012A1
• Filed: 07/24/2013
• Published: 06/02/2016
• Est. Priority Date: 07/24/2013
• Status: Active Grant

• Current Assignee: NEC Corporation, NEC Solution Innovators, Ltd (NEC Corporation)
• Original Assignee: NEC Corporation, NEC Solution Innovators, Ltd (NEC Corporation)
• Inventors: Takahashi, Katsuhiko, MORI, Masateru
• Granted Patent: US 9,646,221 B2
• Time in Patent Office: Days
• CPC Class Codes: G06V 10/48 by mapping characteristic values of the pattern into a parameter space, e.g. Hough transformation
• CPC Class Codes: G06V 10/955 using specific electronic processors

Abstract

A field-programmable gate array (FPGA) coarse Hough transform unit (102) performs on an FPGA a first coarse-precision Hough transform upon an image that has been applied as input to an image input unit (101), and supplies candidate location information obtained by the first Hough transform and the image to an external memory (103). A fine Hough transform unit (104) reads the candidate location information and the image stored in the external memory (103), uses the candidate location information to perform, on a general-purpose processor, a second detailed-precision Hough transform upon the image, and supplies detailed-precision location information of a two-dimensional pattern that is the object of recognition within the image to the external memory (103).

Claims

1. An image recognition apparatus comprising:

image input means that receives an image;

external memory that stores information;

FPGA coarse Hough transform means that performs, on a FPGA (Field-Programmable Gate Array), a first coarse-precision Hough transform upon an image that is received as input in said image input means, and supplies candidate location information that was obtained by said first Hough transform and said image to said external memory; and

fine Hough transform means that operates on a general-purpose processor and that reads said candidate location information and said image that were stored in said external memory, uses the candidate location information to perform a second detailed-precision Hough transform upon the image, and supplies detailed-precision location information for two-dimensional patterns that are the object of recognition in said image to said external memory.

2. The image recognition apparatus as set forth in claim 1, wherein:

said FPGA coarse Hough transform means carries out, as said first Hough transform, a Hough transform for each range that includes a predetermined plurality of pixels in said image, and

said fine Hough transform means carries out, as said second Hough transform, a Hough transform for each pixel of said image.

3. The image recognition apparatus as set forth in claim 1, wherein the voting space included in said FPGA coarse Hough transform means is constituted using a ring buffer.

4. The image recognition apparatus as set forth in claim 3, wherein the number of bin divisions in the vertical direction of the voting space that is included in said FPGA coarse Hough transform means is a value that is greater by 1 or 2 than the maximum vertical size of the two-dimensional pattern in the resolution of said voting space.

5. The image recognition apparatus as set forth in claim 1, wherein said two-dimensional pattern is a traffic sign whose outer shape is round or triangular.

6. A recording medium that records a program for causing a computer to execute procedures of:

receiving an image;

performing, on a FPGA (Field-Programmable Gate Array), a first coarse-precision Hough transform upon said image that was received;

supplying candidate location information that was obtained by said first Hough transform and said image to an external memory; and

causing execution of processes, on a general-purpose processor, of reading said candidate location information and said image that were stored in said external memory, using the candidate location information to perform a second detailed-precision Hough transform upon said image, and supplying detailed-precision location information for two-dimensional patterns that are the objects of recognition in said image to said external memory.