Procedure for using the EdgeLocator step

Draw the search region, if needed. To do so interactively, select this step in the flowchart and click on the Define a region () toolbar button in the Project toolbar. See the Search regions section in Chapter 2: Building a project for more information.

You can tweak the search region's location by specifying its center's coordinates (CenterX and CenterY), size (Height and Width), and Angle (search direction), from the Configuration pane's Search region... button. You can also link the values for the coordinates and size to results from another step to position your search region relative to a previously determined location.

To adjust the edge polarity, smoothing, and number of sub-regions, perform the following:

1. In the Configuration pane, set the Polarity of the edge. A dark to bright transition (positive) indicates an expected rise in grayscale values at the edge in the search direction, whereas a bright to dark transition (negative) indicates an expected decrease in grayscale values at the edge. If you set the polarity to Any, the edge with the sharpest contrast will be located.

2. Set the Edge Fit. Select Best for the strongest edge and Closest for the edge closest to the beginning of the search box that has at least the minimum edge strength entered. The default minimum edge strength is 10.

3. To even out rough edges and remove noise from the image, adjust the value for the Filter Smooth property. Smoothing values can range between 0 (no smoothing) to 100 (very smooth). The default value is 50.

4. If required, increase the Number of sub-regions used to locate an edge in the search region. The EdgeLocator step will find and fit a line to points marking the strongest pixel intensity transition for each sub-region. Increasing the number of sub-regions is useful if the edge is not straight or if its angle varies by 15 degrees or less from the search region's angle. You cannot have fewer than 3 sub-regions (default).

5. 

If the position of an edge is in the right place, however the line is skewed, you should use the edge markers found in the Measurement step, which has more controls available to better identify edges in the presence of noise and unwanted edges. The following images illustrate some examples of situations where you should use the EdgeLocator step, and where you should use the edge markers in the Measurement step. The first image features minimal noise and a clear, simple edge optimally found using the EdgeLocator step. The other images demonstrate an image with more noise. In this case, some noise pixels are interpreted as subedges, and skew the resulting edge marker. To reduce or prevent this from happening, the Measurement step provides maximum sub-position distance controls to limit the effect of outliers.