Find Shapes

Find Shapes is used to locate and measure objects of some particular shape, such as a straight line, a circle, or a search pattern.

Lines

Click the Lines drop-down arrow to see the following options for straight line detection:

Length defines the length of line in pixels. The value assumes that the lines on the original image have a maximum intensity corresponding to the image class (e.g. 255 for 8-bit grayscale). If the lines have lower intensities, the value has to be increased proportionally. Clicking the eye dropper button allows you to select the length right from the image.

When the Find line ends option is on, the application uses the caliper feature to find the exact start and end positions of any detected lines. The longest line along the detected direction is extracted. When the option is off, the lines are drawn between image borders.

The Angles control defines the number of angle steps in the Hough transform (the width of the Hough image).

Filter size is the size of the spike detecting filter. This parameter defines how close spikes on the Hough image could be. If spikes are closer than the filter size, they will be detected as one spike.

Clicking the Create Hough Image button will create a Hough image that can be used to examine the transform using other tools. The Hough image is not displayed in the workspace and only the resulting lines are created.

The Show Demo button opens the demo image and locates the lines.

Circles

Clicking the Circles drop-down arrow displays these options:

Use the Prefilter to select the filter to extract circle edges. Use None if the image is already processed and circles are extracted. Circles must have bright outlines on a dark background.

Min radius and Max radius define the minimum and maximum size of the circles. Clicking the eye dropper button allows you to select the radius directly from the image.

Number of steps defines the number of Hough images to be created by detecting circles from Min to Max radius. This parameter defines the granularity of detected circle radii.

The Completeness parameter defines how complete circles have to be to be located. Set this value to 1 to locate only complete circles. If the value is set to 0.5, then half-circles will be located. The values from 0.6 to 0.8 produce the best results. The Completeness value assumes that the circles on the original image or after edge preprocessing have a maximum intensity corresponding to the image class (e.g. 255 for 8-bit grayscale). If the circles have lower intensities, the value has to be increased proportionally.

Min c-c distance (minimum center to center distance) defines how close circles of the same radius can be. The parameter is in minimal radii. The default value is 1.0, which means that the circles of the same size cannot be closer than minimum radius (Min Radius). If the objects on your image are not overlapping, the parameter should be set to a larger value (e.g. 1.8).

Min radius diff defines the minimum difference in circle diameters when circles have the same center point. The value is relative to the Radius range (Max Radius-Min Radius). The default value is 0.3, which means that the circles can be located separately if their diameters are different for more than 30 percent of the range. If there are no overlapping circles on the image, so that only one circle can be located at the given position, the parameter should be set to 1.

Create Hough image creates an image sequence of circular Hough transforms for all steps.

Show demo button demonstrates the functionality using a demo image.

Irregular Shapes

This button finds a selected pattern on the image. The number of found objects and the locations can be incorporated into a report. Note that search pattern must be set first.

To set the search pattern, draw an AOI around the object you want to find and click the Set Search Pattern button. When the pattern is set, the button is checked and the tooltip shows the current pattern.

The Full FFT checkbox indicates whether or not to use the Full FFT Transform (box is checked) or Phase Only (box is unchecked).

The Objects control sets the maximum number of expected objects. Threshold options define the matching threshold to find all objects.

The Create image button creates an image of cross-correlation between the search pattern and active image.

Clicking the Show demo button demonstrates this feature using a demo image and displays the located cells on an image.