Method of processing image information based on object model

• US 5,887,083 A
• Filed: 07/18/1996
• Issued: 03/23/1999
• Est. Priority Date: 07/19/1995
• Status: Expired due to Term

• Current Assignee: Jiro, President Of Kogyogijutsuin Hiraishi, Stanley Electric Company Limited
• Original Assignee: JIRO Hiraishi President of Kogyogijutsuin Stanley Electric Company Limited
• Inventors: Ishiyama, Yutaka, Tomita, Fumiaki, Sumi, Yasushi
• Primary Examiner(s): Moore, David K.
• Assistant Examiner(s): Chen, Wenpeng
• Application Number:
  US08/683,710
• Time in Patent Office: 978 Days
• Field of Search: 382/151-154, 382/193, 382/199, 382/201, 382/203, 382/206, 382/216, 382/285, 345/419, 345/426, 345/420, 345/430, 345/139
• US Class Current: 382/199
• CPC Class Codes: G06T 2207/10012 Stereo images
• CPC Class Codes: G06T 7/593 from stereo images
• CPC Class Codes: G06T 7/75 involving models
• CPC Class Codes: G06V 10/752 Contour matching
• CPC Class Codes: H04N 13/189 Recording image signals; Reproducing recorded image signals
• CPC Class Codes: H04N 13/194 Transmission of image signals
• CPC Class Codes: H04N 13/239 using two 2D image sensors having a relative position equal to or related to the interocular distance H04N13/243 takes precedence
• CPC Class Codes: H04N 2013/0081 Depth or disparity estimation from stereoscopic image signals
• CPC Class Codes: H04N 2013/0092 Image segmentation from stereoscopic image signals

Abstract

A method of processing image information is disclosed, for recognizing the position and attitude of an object with a high speed treatment and improved recognition accuracy. The method comprised of the steps of entry of a stereoscopic image of an object, extraction of edges of the object, and dividing it into segments based on the local feature of the image, adding the apex information to the segments to produce the recognition data, verifying the recognition data with reference to an object model based on local geometric features to detect a corresponding candidate; finely adjusting each of corresponding candidates based on an entire geometric features; and detecting the position and attitude of the object on the basis of recognition including the initial verification and the fine adjustment.

Claims

1. A method of processing image information by verifying an input image with an object model for determining a position and attitude of an object, comprising the steps of:

entering a stereoscopic image of an object;

extracting edges of said object from said stereoscopic image and dividing said edges into a plurality of segments;

adding apex information to said segments to produce recognition data wherein said apex information is produced through the procedures of;

(1) calculating error information by means of applying straight lines and circles to the segment to be fitted;

(2) correcting said errors by means of regressively dividing said segments into two portions and calculating error data by applying straight lines and circles to the portions;

(3) (a) combining adjacent segments which create straight lines; and

(b) creating tangential lines by combining adjacent segments which is at the end of an arc of a segment which a circle was successfully applied; and

(4) producing an apex having each vector of the tangential lines, said apex of said tangential line being allocated to the middle point of said segment being the minimum distance between the two said tangential lines;

verifying said recognition data with reference to an object model based on local geometric features to detect a corresponding candidate;

finely adjusting each of corresponding candidates on the basis of entire geometric features; and

detecting said position and attitude of said object on the basis of recognition.

2. A method according to claim 1, wherein the step of said fine adjustment includes an initial fine adjustment by merely sample points on every adjacent two segments and a principal fine adjustment by all of said sample points.