Ignore Blobs that Touch the ROI: When enabled, blobs that touch the edge of the ROI will not be detected.
Min Blob Size: Objects smaller than the defined size (in pixels) will be ignored.
Max Blob Size: Objects larger than the defined size (in pixels) will be ignored.
Threshold: The Count Tool’s Blob Count capability displays a histogram of the gray values witin its region of interest. This provides a visual aide when setting the tool’s threshold value. The histogram shows where the majority of the dark pixels (to the left) and the light pixels (to the right) are on the gray scale. You can use the slider to adjust the threshold to a point between the two peaks. The portion of the histogram that is above the threshold is displayed in red; the portion that is below the threshold is displayed in blue.
Selecting Auto Threshold will instruct the software to compute the threshold automatically.
Tolerance: Sets the tolerance on the minimum and maximum number of pixels allowable for the inspection to pass.
Quick Find: Uses a rough location. Faster but less accurate.
Fit Quality: Defines how closely the identified feature must match the template (Relaxed, Normal, Precise).
Accept Threshold: A ratio that determines how well the located feature must match the trained feature. (0.1 – 1, where 1 = perfect match.)
Allowed Rotation: Sets a limit on how much the object can rotate and still be recognized.
Hint: The Count Tool is configured to handle 360 degrees of rotation by default. However, reducing the Allowed Rotation (± Degrees) will speed up the tool's performance considerably.
Tolerance: Sets the tolerance on the minimum and maximum number of pixels allowable for the inspection to pass.
Status – True/False status of the tool.
Count – Number of blobs found.
Status – True/False status of the tool.
Location – Location point (x, y, angle) of highest match shape, normally in pixels. In world-space when calibrated.
Count – Number of blobs found.
Fit Quality – 0 to 1. 1 is a perfect match of object to trained shape.
Effort Level provides a mechanism for adjusting the recognition effort for the tool. There are five options available: Lowest, Low, Normal, High, and Exhaustive. This setting has a large impact on the robustness of the tool at the cost of performance. Normal is adequate for most operations.
Lowest improves speed for easily-recognizable objects on clean backgrounds with mostly-closed contours.
Exhaustive is appropriate for hard-to-recognize objects, consisting of many small features on a cluttered background with mostly discontinuous contours.
Positioning Accuracy provides a mechanism for adjusting the speed/accuracy tradeoff of the positioning process though a scale that ranges from Fast to Accurate. Options are Fast, Coarse, Normal, Fine, and Accurate. Normal is adequate for most operations.
Fast is adequate when accuracy is not as important as performance. This option can be selected for presence/absence applications.
Accurate provides a slight gain in sub-pixel accuracy at the cost of some performance.
Enable Scale Search enables scale tolerance searching to locate the object if it is larger or smaller than a trained model. Note that enabling scale search increases search time, similar to enabling rotation search.
Minimum Scale sets smallest search size of model. A value of 0.5 corresponds to 50%.
Maximum Scale sets largest search size of model. A value of 2 corresponds to 200%.
Conformity Tolerance allows you to set custom conformity tolerance. This is the width of the band +/- in pixels within which the contour is allowed to vary and still be found. If a shape varies slightly between parts, increasing conformity tolerance allows it to be located more easily.
Example: The image below shows a large conformity tolerance. The cross edges can vary anywhere within the yellow boundary and still be considered part of the model.
Min % of Model Match is the value for which a possible match will be further analyzed as a final match candidate. This should always be set lower than Accept Threshold.
Model Name (Read-Only) is the name that AutoVISION automatically assigns to the locate model.