Key Technologies

Real-color Sensing

Real-color processing is an image processing technology that performs high-speed processing of full-color images with a total of 16.7

million colors (256 tones per RGB channel). This means that image processing can be performed with the same color information that is

visible to the human eye, and stable measurements can be performed under lighting that closely resembles natural light.

The camera image is processed as-is without any

loss of quality.

This enables even the slightest of color differences

to be captured with high accuracy.

Captured images are converted to a 256-shade mono-

chrome image and processed. This enables more stable

inspection compared to binary level processing, but

slight changes in color cannot be detected with this

method.

Captured images are converted to a black and white

two-color image and processed. This reduces the

amount of data and enables high-speed process-

ing.

Real color image processing

Color Image Processing

Binary image processing

Previous Image Processing

OMRON FQ2 Series

Black

White

Black

Dark

Bright

Dark

Color variations be-

tween 16.7 million

different colors can

be captured without

any color loss.

17

HDR Sensing

High dynamic range minimizes the

effects of lighting such as halation

and allows highly precise inspections.

Conventional images

Shape Search III (Same functionality included in high-end sensors)

With Shape Search

Ⅲ, you can visualize comparisons between the registered model data the measurement object to easily see when comparisons

are not optimally matched. Visualization of the comparison levels provide the guide for parameter adjustment for acceptable variation and

distortion levels to quickly obtain the best performance. This can save you a lot of time and effort that were previously required.

Even in cases like the following one,

where character registration is required

for image matching methods, no

character registration is required to

read the characters with this new

method, which matches structural

models of characteristic points.

Patent Pending

A A

A

Defects Undetectable Due to Overexposure or

Underexposure

Any spot outside the dynamic range is blurred by halation or shadow.

Dynamic range

of the upper image

Dynamic range

of the lower image

Halation

Underexposure

HDR image

Defects Detectable Even on Reflective or Shadowy Surfaces

The surface of the workpiece is accurately reproduced and detected

even with overexposure or underexposure.

Dynamic range

after HDR

processing

Industry's highest

dynamic range

Max. 16 times

higher than

previous models

New OCR Algorithm: Matching with Structural Models

Structural models record the characteristics of

each character in approximately 80 fonts.

The position and structure of characteristic points are used to

recognize characters.

Background

Changes

Worn

Characters

Size and Font

Changes

Inclined

Characters

Deformation

Chip

Registered model

Measurement image

Reduction in

correlation

No reduction

You can adjust a parameter called

the Acceptable Distortion Level to

enable measurements without

reducing the correlation even if

there is distortion.You can easily

adjust this parameter while

monitoring the comparison.

You can see at a glance the difference

between the registered model and

measurement image.