Edge Detection Measurement

This method extracts parts with color changes as edges to perform measurement. Edges are found through color changes in the measurement region.
It finds edges using color changes in the measurement region.
Illustration of edge measurement

Edge Level

Indicates the edge color change level (degree of color difference). This level is adjusted if edges cannot be accurately detected.
Note
  • The value 0 to 100 for the edge level indicates the edge intensity.It is not related to color differences in the original image.
For case of measurement using relative position (%) with regards to width of color difference
The edges are detected in the following manner.
  1. Calculate the overall distribution of color difference in the measurement region.
  2. Min. color difference value: 0%; Max. color difference value: 100%
  3. Locations in which there is an edge level color difference are detected as edges.
    Edge Position
    Illustration of edge level
    Edge Pitch
    Illustration of edge level
When performing measurement using color difference value
Edge level is set using color difference absolute value.
Illustration of edge level

Noise Level

This level judges whether an edge is present or not.
The maximum value and minimum value for color difference in the edge detection area are determined and if the difference is less than the noise level, it is judged that there is no edge. When detection is affected by noise, increase this value.
(within area)
Maximum value - minimum value of color difference < noise level - edge does not exist - measurement result is fail
Maximum value - minimum value of color difference > = Noise level - edge exists - targeted for measurement
Example: When noise level is set to 30
Illustration of noise level

Noise Width

Set the width for judging noise.
  • When no color is specified
    If another edge is detected within the width range starting from the point where an edge was first detected, the newly detected point is considered noise.
  • When a color is specified
    If the color difference distributions again falls below the edge level within the width range starting from the point where an edge was first detected, the newly detected point is considered noise.
Illustration of noise width
When points are mistakenly detected as edges due to noise, increase the color difference value.