Component shape recognition method

• US 4,981,372 A
• Filed: 10/21/1988
• Issued: 01/01/1991
• Est. Priority Date: 10/23/1987
• Status: Expired due to Term

• Current Assignee: Matsushita Electric Industrial Company Limited (Panasonic Holdings Corporation)
• Original Assignee: Matsushita Electric Industrial Company Limited (Panasonic Holdings Corporation)
• Inventors: Mohri, Akira, Ohkawara, Daisuke, Okumura, Kazumasa, Morimoto, Masamichi
• Primary Examiner(s): Boudreau, Leo H.
• Assistant Examiner(s): Santos, Daniel
• Application Number:
  US07/260,998
• Time in Patent Office: 802 Days
• Field of Search: 382/8, 358/107, 356/376, 356/375
• US Class Current: 382/151
• CPC Class Codes: G06V 10/24 Aligning, centring, orientation detection or correction of the image
• CPC Class Codes: G06V 10/46 Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features colour feature extraction G06V10/56

Abstract

A boundary of leads of an integrated circuit component shown by binary picture elements on a recognition plane of an image display means is detected by a conventional boundary detecting method, and respective corners of the leads are detected on the basis of variation of an angle between a predetermined axis and a line connecting two picture elements having a predetermined distance from each other on the boundary. Moreover, a shape of the lead is recognized on the basis of the shapes of the corners. Subsequently, a search zone containing a tip portion of one lead is set by referring the midpoint of the tip of the lead, and is scanned by horizonal scanning lines and vertical scanning lines. Then segments of both the scanning lines overlapping on the tip portion are obtained, and a protruding direction of the lead is detected referring to a segment touching on a boundary of a search zone. Furthermore, a midpoint of the tip is calculated from coordinate values of plural segments crossing the tip portion.

Claims

1. A component shape recognition method comprising the steps of:

detecting a boundary of an object on a recognition plane of an image display means,

detecting corners of said object on the basis of variation of an angle between a predetermined axis and a line connecting two picture elements having a predetermined distance on said boundary,

detecting a shape of said object on the basis of positions and shapes of said corners,

detecting a midpoint of a tip portion of said object,

setting a search zone containing said tip portion of said object on said recognition plane, to be scanned by horizontal scanning lines and vertical scanning lines,

calculating segments of said horizontal scanning lines and said vertical lines overlapping with said object on the basis of the respective coordinate values,

creating data of positions of said tip portion of coordinate values on said horizontal scanning lines and coordinate values on said vertical scanning lines,

detecting a protruded direction of said tip portion on the basis of a segment touching on a boundary of said search zone, and

detecting a midpoint in a cross direction of said tip portion by calculating an average of said coordinate values of plural segments crossing said tip portion.