ViewHidden Preview TextV

View in your browser



Innovation with Integrity

In this issue

▪ Method note MCT-075 – Watershed Separation

▪ Image of the month

▪ Bruker microCT news

▪ Upcoming events


Method Note MCT-075 – Watershed Separation

When analyzing grain-based materials such as rocks and sediments, pellets, pharmaceutical particles, powders, etc., the properties of individual particles are often of interest. However, since these particles are usually stacked on top of each other, getting these individual properties is often a problem. When performing a basic segmentation, the stack of particles usually forms just one big cluster, appearing as one single object. For quantification of each individual object in terms of density or morphometric parameters such as volume, surface, x, y, z location,… this is not desirable. Two pores, particles, fibers,… which are touching will be counted as one, leading to miscalculations.

A number of despeckle and morphological operations are provided in CTAn to deal with such issues, one of them is the watershed separation.

The watershed separation algorithm aims to divide connecting objects into separate objects. The name comes from geographical watershed lines, dividing landscapes and continents in individual basins, where mountain ranges are considered as separation lines (watersheds) between these basins.

The algorithm is composed of several steps:

•    Segmentation
•    Watershed separation
      o    Calculation of distance map
      o    Defining (local) minima in the distance map
      o    Separation into individual objects.

The first step in performing a watershed separation is the segmentation step (figure 1). Previous method notes (including MN059 – Porosity Analysis, MN020 – Adaptive Thresholding, MN011 – Advanced Porosity Analysis) describe multiple ways to perform this step.


Figure 1 (left) original image; (right) segmented image


The first part of the watershed separation algorithm is to transform the dataset into a distance map (figure 2). For each point within the object, the shortest distance to the edge of that object is calculated. A distance map is thus an image where the grey value is related to the distance to the edge of the object. This process step happens automatically in the background.

Figure 2 (left) segmented image; (right) distance map


Figure 3 (left) original binary image; (right) result of the watershed separation, performed in 3D space


The distance map will be used as an input parameter to separate the binary input dataset. Each of the individual particles is separated from the neighboring particle by a thin black line, just one voxel wide in areas where the particles were touching (figure 3).

The watershed algorithm in its default state is very sensitive to local minima. Therefore, CTAn includes a tolerance value to ignore local minima and avoid over-separation.

The complete workflow of the watershed separation is described in Method Note 075 – Watershed Separation, and illustrated on a dataset of pharmaceutical particles.

Image of the Month

A sandstone rock scanned with SKYSCAN 2214 at 2.1 micron/voxel. The very sharp image quality over the whole field of view of 6.4 mm by 6.4 mm enables to clearly distinguish the different minerals.

* Courtesy of Shell


Volume 6, Issue 4
August 2019







Bruker microCT news


- BrukerSupport: Software Updates & Method Notes

We have pre-registered our users at for the distribution of documentation and software. Please log into your account and select your SKYSCAN microCT systems.

At you'll find the latest software downloads available for you. In addition, the complete library of method notes has been uploaded as well.

If you did not receive an email for log-in, please apply for an account with your system serial number. Enjoy your support membership and let us know if you have any questions!




Upcoming events

Bruker microCT will participate with an exhibit in these upcoming conferences. Please click the links below for more information. We hope to see you there!


August 26 - 30, 2019
SCA, The Society of Core Analysts, Pau, France


September 04 - 07, 2019
WMIC, The World Molecular Imaging Congress, Montréal, QC, Canada


September 04 -11, 2019
Volume Graphics User Group Meeting, Heidelberg, Germany


September 20 - 23, 2019
ASBMR, The American Society for Bone and Mineral Research, Orlando, FL, US


September 29 - October 3, 2019
MS&T, Materials Science & Technology, Portland, US


November 19 - 22, 2019
FormNext, Frankfurt, Germany




You have been sent this email as a valuable contact of Bruker microCT. We take your privacy seriously and therefore updated our Privacy & Cookies Policy. You may unsubscribe from our emails at any time. 

© Copyright 2019 Bruker microCT, a division of Bruker Corporation. All rights reserved.

Bruker microCT | Kartuizersweg 32550 Kontich | Belgium