Introduction to Dynamic Image Analysis

© 2010 HORIBA, Ltd. All rights reserved.

Introduction to Dynamic 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 resultsReplace sieves for size distribution analysisDecide whether you have single particles or agglomeratesSeeing is believing

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


Why Image Analysis

Need shape information for evaluating•Packing•Flow•Tendency to create dust•Abrasive performance•Optical properties (e.g., reflectivity)


Why Dynamic Image Analysis?Robust measurement….the interaction

between the instrument and the particle is optical, so there is no wear and change in calibration.

High resolution size distribution resultsFast

Also, these are all reasons to use Dynamic Image Analysis instead of sieves.


Weigh % sample caught on known screen sizesSolid particles 20 μm – 125 mmLow equipment costDirect measurement methodSome automation/calculation available

More information available through www.retsch.com

Sieves


SievesTend to wear over timeDifficult to tell when sieve results are “drifting” due to wearResults depend on nature of “shaking” leading to operator to operator variations in results.

Rotap or Vibratory or Manual?Sieve overloading (too much material)intervals?How long do you shake?Amplitude?

Limited informationFinite number of sievesNo shape informationCubes and needles interact in unexpected ways with the sieve.


Dynamic:particles flow past camera

1 – 3000 um

Static:particles fixed on slide,stage moves slide

0.5 – 1000 um2000 um w/1.25 objective

Two Approaches


Major Steps in Image Analysis

Image Acquisitionand enhancement

Object/Phasedetection

Measurements


Features

Use gravity, or, better, vacuum (from a compressed air supply and venturi) in order to draw particles through instrument. Vacuum helps keep the windows clean.

9© Retsch Technology GmbH

Dynamic Image Analysis:Moving Particles


Acquiring ImagesWe want a good microscope and nice sharp

images.Pay attention to lighting and focus.

No Yes


wider depth of focal plane

sharper images

= more resolution

weaker light source => wider aperture

stronger light source => smaller aperture => better images

blurry images

= bad resolution

More Light better Sharpness



Pixel =

Picture Element

Height of fallduring capturing time

Slower (100 µs)

due to weak light source.

CCD-Basicpixel raster

CCD-Zoompixel raster 20 µm

Image Capture Speed (Time)



Height of fallduring illumination time

CCD-Basicpixel raster

CCD-Zoompixel raster 15 µm


Capturing time influence

Bright light source (and sensitive CCD) means no blurring.


Dispersing a SampleWant to spread particles out so that they

don’t touch.No Yes


Want to spread particles out so that they don’t touch.

Use % of field of view that is covered in order to control feed rate. Try 1% at first.

Control feed rate.

Feeding Too fast Good


CCD - Basic CCD - Zoom

ResolutionMeasuring Principle



Two-Camera-SystemMeasuring Principle

Basic

Zoom

Basic-Camera Zoom-Camera



ResolutionMeasurement Results

mixture of six sizes of grinding balls18© Retsch Technology GmbH


comparison

CAMSIZER-measurement xarea (red)and sieving * (blue)

Competing Measuring Methods

A

A‘ = A

x are

a

xarea“diameter

via projection surface”

Digital Image ProcessingArea Measurement Sieving



x [mm]0.1 10

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Q3

Tinovetin-B-CA584A_BZ_xc_min_002.rdfSyngenta-1mm-2min-Sieb.ref


--- width measurement

-*- Sieving

comparison

CAMSIZER-measurement xc min (red)and sieving * (black)

xcmin

xc min

“width”

Digital Image ProcessingMeasuring of Width Sieving




x [mm]1.0 1.25 1.5 1.75 2.00

10

20

30

40

50

60

70

80

Q3 [%]

rice

xcmin

A‘ = A

x are

a

A

xc min (width) is more similar to sieve result than xarea

CAMSIZER Sieving



x [mm]0.2 0.4 0.6 10

10

20

30

40

50

60

70

80

Q3 [%]

Sample A_BZ_0.2%_xc_min_001.rdfSample A_.ref

Digital Imaging Sieving




fitted result

CAMSIZER-measurement xarea (red)to sieving * (blue)


Fitting of CAMSIZER results to Sieving



x [µm]200 400 600 8000

10

20

30

40

50

60

70

80

90

Q3 [%]

RT669_3993_Z_LB_05%_xc_min_001.rdfRT669_RT_3993.ref

Examples of samples with lens shaped particles without fitting

CAMSIZER-result xc min (red)sieve analysis * (black)





xc_min [mm]1 2 3 40

0.1

0.2

0.3

0.4

0.5

0.6

0.7

q3 [1/mm]

b/l0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.90

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Q3

New elementary fitting with single (narrow) sieve class

Two samples with similar shape




Use CAMSIZER data from a single sieve class (an “element” of the distribution” to provide superior fitting.


xc_min [mm]1.0 1.5 2.0 2.5 3.00

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Q3

xc_min [mm]1.0 1.5 2.0 2.5 3.00

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Q3

xc_min [mm]1.0 1.5 2.0 2.5 3.00

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Q3

Q3 – Fitting does not fit both samples

Elementary - Fitting

works for both samples





Feret diam. 1

Feret di am. 2

Longest diam.

to longest┴

Equivalentsphericaldiam

Size Descriptors


Feret diam. 1

Feret diam. 2

Longest 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 the deviation between results from different measurements. It 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

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.


More particles means more accuracy.

5000

And, it’s really easy to flow a lot of particles through a dynamic image analyzer!


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.


Measurement of very broad particle distributions (due to speed and two cameras)Direct particle definition

by width (analogue to sieving)by lengthor projection surface

Two camera system for more accuracy and reproducibilityEasy operationFail-safe, robustIdeal for particle shape analysesMeasurement of density, counting of particles

Measurement Results


CAMSIZER Advantages


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


Watch out for:

Sample preparationImage qualityMeasure enough particles


Dynamic Image Analysis is good for:

Replacing SievesSizeShapeSupplementing other techniques


Questions?

www.horiba.com/us/particle

Jeffrey Bodycomb, [email protected]

732-648-3431