Skeletonisation to Find the Centre of Dendrites Traced from a 2D Microstructural Image

Joshua Miller, Martin Strangwood, Sonja Steinbach, Nils Warnken

Research output: Chapter in Book/Report/Conference proceedingConference contribution


As three-dimensional analysis of dendritic microstructures (such as Ni-superalloys) has become more prevalent with the use of techniques such as Robomet, synchrotron and computerised tomography, there has been a greater need for accurate and efficient methods to characterise dendrites in three dimensions. One parameter which is needed for a variety of further calculations is the primary dendrite arm spacing (λ_1). This parameter can be calculated by finding the distances between all of the centres of the dendrites on a single plane. However before this can be computed the locations of all these centres must be obtained. One technique which has often been used in the literature is the centre of geometry method, which simply determines the centre of gravity of each dendrite as seen in a micrograph. However, this method has some major drawbacks. For instance it can assign the centre outside of a dendrite, which often applies to irregular shaped dendrites. As a result another technique shall be presented using skeletonisation. It is shown here that this technique represents good improvement over the centre of geometry method; however it also has drawbacks in certain specific conditions. Combined use of all these methods however shows promise in the search for a reliable method to identify the centres of differently shaped dendrites.
Original languageEnglish
Title of host publicationSolidification Processing 2017
Subtitle of host publicationProceedings of the 6th Decennial International Conference on Solidification Processing
EditorsZhongyun Fan
PublisherBrunel University
Number of pages4
ISBN (Print)978-1908549297
Publication statusPublished - 25 Jul 2017


  • Dendrite Centre
  • Skeletonisation
  • Image Analysis
  • Microscopy


Dive into the research topics of 'Skeletonisation to Find the Centre of Dendrites Traced from a 2D Microstructural Image'. Together they form a unique fingerprint.

Cite this