introduction to static image analysis

© 2010 HORIBA, Ltd. All rights reserved.

Introduction to Static Image Analysis

Jeffrey Bodycomb, Ph.D.HORIBA Scientific

www.horiba.com/us/particle


Why Image Analysis?Need shape information, for example due to importance of powder flowVerify/Supplement diffraction results

These may have the same size (cross section), but behave very differently.


Why Image Analysis?Crystalline, acicular powders needs more than “equivalent diameter”

We want to characterize a needle by the length (or better, length and width).


Why Image AnalysisPictures: contaminants, identification, degree of agglomerationScreen excipients, full morphologyRoot cause of error (tablet batches), combined w/other techniquesReplace manual microscopy


Why Image Analysis

Need shape information for evaluating packing and flow.


Effect of Shape on FlowYes, I assumed density doesn’t matter.Roundness is a measure based on particle perimeter.

θc


Major Steps in Image Analysis

Image Acquisitionand enhancement

Object/Phasedetection

Measurements


Dynamic:particles flow past camera

Static:particles fixed on slide,stage moves slide

1 – 3000 um0.5 – 1000 um2000 um w/1.25 objective

Two Approaches


Dispersing a SampleWant to spread particles out so that they

don’t touch (too much).No Yes


Acquiring ImagesWe want a good microscope and nice sharp

images.Pay attention to lighting and focus.

No Yes


CLEMEX

Stack images for sharper final image

Out of focus

Multilayer Grab for SharpnessIn focus


Image Binarization

Turn into binary image (i.e., decide what is a particle and what isn’t).


Separation


Long narrow (acicular) particles tend to touch a lot. So, this is a typical problem.

Fiber Separation


Fiber Separation

Separate and assign each fiber to its own bitplane.


Fiber Separation

Now we are ready for analysis.


SortingRoughness, “too rough” is red

Roundness, not round enough is green


Statistics

Only a few particles for this example, so the result is not useful.


Feret diam. 1

Feret di am. 2

Longest diam.

Shortest diam.to longest┴

Equivalentsphericaldiam

Size Descriptors


Feret diam. 1

Feret diam. 2

Longest diam.

Shortest diam.to longest┴

Aspect ratio

= shortest diamlongest diam

= to longest diamlongest diam

= shortest Feret diamlongest Feret diam

= three different numbers!

┴

Shape: Aspect Ratio


More Shape Descriptors


http://www.spcpress.com/pdf/Manufacturing_Specification.pdf, By David Wheeler

Must tighten internal spec by lab error %Then product always within performance specification

Specification with Measurement Error

Allowance for measurement uncertainty

Allowance for measurement uncertainty


20,000 particles 200 particles

second populationmissed

“holes” in distribution

But d10, d50 &d90 may appear similar

Effect of Number of Particles Counted


Two Kinds of Standard Deviation!

Sample standard deviation is a property of the sample. It is the width of the size distribution.Measurement standard deviation is a result of the measurement and is affected by the sample standard deviation.


More particles for more accuracy.

0

1020

3040

5060

7080

90

100

200

300

500

1000

2000

5000

1000

020

000

5000

010

0000

1771

87

dv10 dv50 dv90 dv10 dv50 dv90100 28.004 41.983 56.618200 29.551 42.914 58.466300 28.223 43.722 65.737500 29.891 45.953 79.187

1000 29.729 46.826 79.2182000 31.292 47.899 79.3785000 30.948 47.463 81.923

10000 31.433 48.57 81.82220000 31.662 48.992 81.99850000 31.826 49.157 83.435

100000 32.381 49.766 84.601177187 32.742 49.833 83.873

Assume 49.833 is “correct”dv50,0.95 x 49.833 = 47.34 is within 95%,47.463 achieved by 5000 counts!

5000

Use this to control precision of your data (and not spend extra time on precision you don’t need.


Divide Large Data Set into Smaller Sets

Error bars are one standard deviation from repeated measurements of the same number of particles from different parts of the sample. The error bars get smaller as you evaluate more particles.


The HORIBA PSA300

Turnkey SystemMore time getting results and less time engineering

AutomatedFasterLess operator laborLess operator bias

Powerful Software FeaturesImage EnhancementParticle separation

Separate Disperser OptionMore flexible sample preparation


Static or Dynamic Image Analysis?

DynamicBroad size distributions (since it is easier to obtain data from a lot of particles)Samples that flow easily (since they must be dropped in front of camera)Powders, pellets, granules

StaticSamples that are more difficult to disperse (there are more methods for dispersing the samples)Samples that are more delicatePastes, sticky particles, suspensions


Conclusions

Image Analysis is good for SizeShapeSupplementing other techniques

Watch out forSample preparationImage qualityMeasure enough particles


Questions?

www.horiba.com/us/particle

Jeffrey Bodycomb, [email protected]

866-562-4698

introduction to static image analysis

Technology

horiba scientific

horiba psa300turnkey

sharperfinal image

image analysispictures

binary image

image binarization

number of particles

particles counted20