Using captured images, (our software) is able to determine position coordinates or the work, presence or absence of work details, and whether the inspected work is OK or NG.
Image processing inspections employ the five kinds of basic processing as shown below.
(D/P+GIGI) and the condition of the work is inspected during each Inspection process.
Image processing technology is an industrial field that is continually being researched and developed as a replacement for visual inspections that were previously done using the human eye.
The range of applications continues to grow in tandem with the progress of technology.

Alignment

Guidance is provided on where objects of inspection are by using robotic eyes.

Normalized Correlation

This search method calculates the correlation coefficient (degree of similarity) between part of an image and a template and then uses that similarity as a yardstick.

Geometric
Pattern Matching

Features (geometricinformation) are extracted from an image to specify the position.

Inspections

Defects are checked for based
on a comparison with non-defective products.

DefFinder®

An inspection method developed by our company that makes it possible to detect scratches, dirt, foreign matter, and other defects by comparing image differences.

Scratch / Foreign Matter Detection Filter

A detection filter that extracts scratches and cracks from plain parts.

Measurement

Even if the object is not in the form of straight lines, noncontact measurement can be performed.

Edge Detection

Edge peaks are extracted by differentiating objects of inspection in an image.

Recognition

Recognize 2D code,
characters, and so on

2D Code Reading

2D codes (e.g. data matrix, QR codes, etc.) in an image are searched for and then read as encoded character strings.

Presence/direction

Does the object of inspection exist？
Which way is it facing？

Blob Analysis

This is a method during which a gray image is used as a threshold reference and an analysis is performed after binary conversion.