Mathematical Morphology

The PYXIS DGGS SDK includes Dilation, Erosion, Open, Close mathematical morphology processes.  

Mathematical morphology is the analysis of a signal as a shape, and works by changing the shape within the signal. It is a computationally efficient technique for manipulating the shape of a signal or image, and is commonly used for image enhancement and restoration, texture analysis, and noise reduction. In the PYXIS DGGS SDK, a Morphological Transformation is applied to every cell of an image, and produces a result based on the value of that cell, and the value of its neighbors. 

In an Erosion, the cells output is true only if the cell and all of its neighbors are true, thereby “eroding” the edges of shapes away and removing smaller shapes entirely. In a Dilation, a cells output is true if any of the source cells are true, causing all shapes to grow, and nearby shapes to merge. The Open operation builds on these basic morphologies with an Erode and then a Dilate of the input data in order to remove small shapes from the image completely while leaving larger shapes unchanged. Likewise the Close operator first Dilates then Erodes the signal to removes small gaps in the signal while leaving the rest unchanged.