My watch list  

3D PartAn: A new generation of DIA Dynamic Image Analysis Systems

Dynamic Image Analysis much closer to reality now

Robert Waggeling

The 3D technique is realized in the model PartAn 3001 analyzer. It can be used for laboratory and process applications. In an impressive way, 3D can be opposed to 2D in one and the same instrument. The measurement comfort is the same in both techniques; however the benefit of 3D is incomparably higher.

Mile stones of DIA Dynamic Image Analysis

2D - DIA

“2D” stands for making one image per particle during its passage through the optical compartment. It is the common technique in most commercial instruments.

In 1987, the PartAn/DIA System was born as an online particle size and shape analyzer for fertilizer plants. It was supported under DOS software reading out 3 images per second. As it was the first dynamic image analyzer, the system was regarded as the pioneer for a new age in image analysis.  As new applications came up, a laboratory version was developed known today as the PartAn 2001L. To respond to the need of customers for laboratory automation, an Auto Sampler was developed.

3D - DIA

“3D” stands for making several images per particle during its fall through the optical compartment, hence seeing it in different orientations.

With the faster PC generation it was possible to acquire the data from high speed cameras, do the calculation in real time and display the results quasi-simultaneously. The multi-imaging technique was born and patented. For simplicity, we call it the “3D” DIA dynamic image analysis. In 2010, the first systems where installed.  3D does neither require more sophistication from the user, nor the user has to resign comfort.  3D DIA will be the standard future technique.  It is described hereafter.

With a frequency up to 120 images per seconds all particles are analyzed 6 to 12 times in real time putting one particle in various orientations.  Projected on a high resolution camera and analyzed online, these images deliver revolutionary results.   

Conventional dynamic image analyzers such as the foregoing 2D PartAn, just use one image per particle. Even analyzers with more than one camera do not reach the same resolution that the 3D PartAn technology offers.

Figure 1: Macaroni measured by a caliper.

3D versus 2D: Analysis examples

As the PartAn can be operated in 3D as well as in 2D mode, the difference can be shown very impressively.

Length analysis

In the example of Macaroni the superiority of the 3D technique for length measurements is obvious.

Figure 2: Results 2D - versus 3D –analysis method: length measurement

The easiest way to compare the length measurement is to use a caliper (Fig. 1) and compare the particle length result with the DIA, once measured with 2D and once with 3D. Measured with the caliper, the particle length of the Macaroni is between 9.3 and 11.8 mm. In Fig. 2, the 2D (read and green curve) and 3D (black curve) length distributions are confronted.  The 2D analysis shows a length distribution between 5.4 and 12 mm, with an unreal tail to short lengths. In 3D the particles where measured 8-times in different orientations. Statistically, the real length appears more often than in 2D. Accordingly, it is better weighted. The result is a distribution between 9 and 12 mm, which comes very near to reality.  Interestingly, the 2D result does not show any difference between 25 and 100 images per second. That proves that the difference in accuracy between 3D and 2D cannot be explained by statistics. 3D is simply more realistic.

Figure 3: Example Macaroni: Thickness distribution 2D – versus 3D – Dynamic Image Analysis

Thickness analysis

If we look at the thickness distribution we get a particle size shift into the opposite direction.  The average of the 2D thickness measurement gives a coarser result then the exactly analyzed and manually measured real particle thickness (Fig. 3). The difference here is less impressive, but still significant.

Figure 4: Free falling coins in different orientations.

The next example (Fig. 4) shows the 3D measurement of free falling coins. The coins fall from left to the right and rotate on the way down. As we see here in this example, the rotation is not always the same. Even the 3D analysis doesn’t show the exact thickness of each particle.  However, the 2D calculation of the length, thickness, min/max area and min/max perimeter is based on the measurement of one image and one orientation per particle.

Figure 5: Granulate shown in different positions.

2D/3D – Comparison: Length and thickness of a granulate

Min: 2,056 1,976 2,163 2,847 2,526 2,055 2,409
Max: 3,444 4,291 4,484 4,149 3,715 3,877 4,436
MaxL: 3,355 4,381 4,631 4,224 3,645 3,903 4,467

Min 3D = 1,976  Max 3D = 4,484 MaxL 3D = 4,63

Figure 6: length distribution of granulates, 2D versus 3D.

There are many more examples that show the significant benefit of the 3D method in comparison to other dynamic image analyzers.

Figure 7: Thickness distribution of granulates, 2D versus 3D.


For the same time and labor, the result of the 3D Dynamic image analysis s more accurate.  The current excellent response of the worldwide market confirms that this technology is rapidly growing.  Application examples in research and quality control are available in abundance. Systems are working online in quarries, fertilizer production, food processing and steel plants with measurement ranges up to 120 mm.

More about Microtrac
Your browser is not current. Microsoft Internet Explorer 6.0 does not support some functions on Chemie.DE