Edge Measurement

This measurement method will extract the parts with color changed as edges, and find out edges according to color change in the measurement region.
Illustration of Edge Measurement

Edge Level

It indicates the color change level of edges (degree of color difference). This level shall be regulated in the case of failure to detect the edge.
The edges are detected using the following methods:
  1. Calculate overall distribution of color difference in the measurement region.
  2. Min. specified value of color difference: 0%; max. specified value: 100%.
  3. Edge level with change of color difference is detected as edge.
    Edge Position
    Illustration of Edge Position
    Edge Pitch
    Illustration of Edge Pitch
The edge level shall be adjusted when edge cannot be found due to low color difference of background, or noise shall be removed.
Note
  • Edge level 0 to 100 represents "strength of edge", irrelative with color difference of original images.

Noise Level

Used to judge the level of edge.
When calculating max. and min. value of color difference in the "Edge detection" area, no edge exists if the difference is less than noise level. When the detection is affected by noise, increase the value.
(within area).
Maximum value of color difference - minimum value < Noise level Icon judging as absence of edge Icon measurement results: NG.
Maximum value of color difference - minimum value >= Noise level Icon judging as presence of edge Icon taken as measurement objects
Example: when noise level is set as 30
Illustration of noise levels

Noise Width

Set the width of noise.
  • When no color is selected
    When the edge Error is detected again, within the range from the location of edge to noise width during first detection, the detected points are considered as noise.
  • When specifying color
    When the distribution of color difference is less than Edge level again, within the range from the location of edge to noise width during first detection, the detected points are considered as noise.
Illustration of noise width
When the detection is affected by noise, increase the value.