Morphology

Toolbar:

Menu:    Processing > Morphology…

Use this command to apply morphological operations to an image. Like spatial filters, morphological operations also work with pixel neighborhood and kernels, however they do not multiply the neighborhood values by kernel coefficients. Instead, morphological operations use comparison and counting to modify the pixel upon which it is focused. The kernel is moved across the image pixel by pixel. Where the active elements of the probe match the image, no change occurs. Where the intensity of the image pixel is different than that of the kernel, the image pixel is altered according to the morphological operation selected.

Morphological functions use different approach when they are applied to different types of images. When the image is of a multiphase type, the binary morphology operations are used. For that ImageWarp uses a special technique that separately processes pixels belonging to different phases. When the image is of a gray type including high bit depth, the gray morphology operations are used. Finally, if the image is a color one, ImageWarp employs the color morphology, performing operations in certain color space defined by the Processing space option in the Preferences If the color space for processing differs from the input image type, ImageWarp converts the image into the corresponding internal type, processes it and then converts it back into the original type.

Morphological functions fall into one of four general categories: basic morphology, geodesy, connectivity and user-defined kernel operations. Each category is presented in a separate tab of the Morphology functional dialog box. Select one of the following functions:

Basic

Click… To… Erosion erode the bright objects and dilate the dark ones. Dilation dilate the bright objects and erode the dark ones. Opening separate narrowly connected bright objects and remove small ones. Closing fill gaps in bright objects and connect objects that are close together. Contours extract contours of objects. Thinning reduce objects to their skeletons. Thickening grow objects around their skeletons. Pruning remove branches from skeletons. TopHat to detect small bright regions in an image. BottomHat to detect small dark regions in an image. Convex build convex hulls of objects in a multiphase image.

Geodesy

Click… To… Distance map creates a distance map of an image. Direction map create a direction map of an image. Local minima find local minima of intensity in an image. Local maxima find local maxima of intensity in an image. End points find end points of skeletons or ridges. Node points find node points of skeleton or ridges. Saddle points find saddle points of skeleton or ridges. Ultimate erosion reduce objects to single pixels. MedialAxis perform medial axis transform on a multiphase image. Separation separate touching convex objects in a multiphase image. Watershed find ridges of intensity in an image. Waterfall create basins of intensity in an image.

Connvectivity   

Click… To… Identify to identify connected components in a binary image. Label to label or erase an object in a multiphase image. Fill holes to fill holes in objects in a multiphase image. Border kill to remove objects touching boundaries in a multiphase image. Feature AND to extract the features that are "hit" by pixels in a mask image. Scrap to reject or select objects in a certain range of sizes. Skip To reject or select objects based on a supplied parameter range. Classify To classify objects based on a supplied parameter. Link lines to link broken lines in a multiphase image.

Kernel

Click… To… Erosion with kernel perform erosion with a user-defined kernel. Dilation with kernel perform dilation with a user-defined kernel. Opening with kernel perform opening with a user-defined kernel. Closing with kernel perform closing with a user-defined kernel. TopHat with kernel apply tophat operator with a user-defined kernels. BottomHat with kernel apply bottomhat operator with a user-defined kernel. Template perform template matching on a multiphase image. Hit Or Miss apply "hit or miss" operator with two user-defined kernels.