Stabilizing Measurement

This section describes key points for adjustment when measurement is not stable. There are two methods for improving measurement precision: Performing processing of images loaded from the camera (filtering) or adjusting settings and parameters.

Adjusting Parameters of Each Processing Item

Adjustment to improve precision and stability varies depending on the processing item.

For details, see "Key Points for Adjustment" for each processing item in the Processing Item List Manual.

Processing Images (Filtering)

There are cases in which high-precision measurement is impossible such as when using images loaded from the camera that have noise, irregularities, or low contrast or when the background has patterns during defect measurement. In this case, measurement accuracy can be improved by performing processing of measurement images in advance.

"Processing Item List Manual", "Filtering"

When measurement images have irregularities (search and location positioning are not stable)

The filtering items "Smoothing (strong and weak)" and "Median" are both effective.

Smoothing (strong and weak)
This processing changes the shade of images so that irregularities are not as easily seen.
Median
In comparison with smoothing, "Median" allows for irregularities to be hidden without having to shade the edges of images.

When measurement images contain noise

The filtering items "Dilate" and "Erosion" are both effective.

Dilate
When there is dark noise in an image, bright areas are enlarged to eliminate dark noise.
Erosion
When there is bright noise in an image, bright areas are contracted to eliminate bright noise.

When contrast of measurement images is low (defect inspection is unstable)

The filtering items "Extract vertical edges", "Extract horizontal edges", and "Extract edges" are effective.

Extract vertical edges
This extracts the vertical edges of an image.
Extract horizontal edges
This extracts the horizontal edges of an image.
Extract edges
This extracts the all edges of an image.

When unidentifiable shapes are present

The filtering item "Extract edges" is effective.

Extract edges
This is used to make the profile clearer and the shape more identifiable.