Browse Prior Art Database

Multiple-Image Vision Inspection Process

IP.com Disclosure Number: IPCOM000040049D
Original Publication Date: 1987-Sep-01
Included in the Prior Art Database: 2005-Feb-01
Document File: 3 page(s) / 69K

Publishing Venue

IBM

Related People

Eng, RC: AUTHOR

Abstract

This article describes a method to inspect items by capturing several images of the item under different external conditions and combining them mathematically to increase the amount of information discerned by the vision system. Vision inspection systems minimally consist of a vision processor, at least one video camera, and a subject item to be inspected. The item is to be inspected by the vision system, and the system will determine whether the item is good or defective. Current vision inspection systems operate by focusing the video camera on an area of interest from the item. A single image of this area is captured by the camera, digitized, and then processed by the vision processor. A single image plane is used for each area to be inspected of the item.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 54% of the total text.

Page 1 of 3

Multiple-Image Vision Inspection Process

This article describes a method to inspect items by capturing several images of the item under different external conditions and combining them mathematically to increase the amount of information discerned by the vision system. Vision inspection systems minimally consist of a vision processor, at least one video camera, and a subject item to be inspected. The item is to be inspected by the vision system, and the system will determine whether the item is good or defective. Current vision inspection systems operate by focusing the video camera on an area of interest from the item. A single image of this area is captured by the camera, digitized, and then processed by the vision processor. A single image plane is used for each area to be inspected of the item.

(Image Omitted)

The vision inspection method disclosed herein uses multiple (two or more)-image planes for each area of the subject item to be inspected. Several image planes (frames) of the same area are captured by the camera and digitized. However, the conditions under which each image plane (frame) is captured is different; for example, the lighting of the item is changed between image planes, or color filters on the camera lens are changed between image planes. The multiple image frames which are captured are images of exactly the same area and objects of the item. The camera and item are not moved between the collection of the multiple images. However, each image plane represents a picture of the item under different lighting, color, etc., conditions. The vision inspection system's processor then performs one or more mathematical operations using the multiple-image planes as operands. The resultant is a single-image plane from which the results of the item inspection can be determined (good or bad item). The multiple-image inspection system consists of the subject item to be inspected, one or more video cameras, a vision processor, and a controller which can electrically or mechanically change the conditions under which an image is captured by the camera, such as lighting, light filtering...