Aditya MSc Start Seminar

1

In-line laser light backscattering image analysis to monitor physico-chemical

properties of ginger during drying

M.Sc. ThesisAditya Parmar

Supervised by:Dr. Giuseppe Romano

Dr. Marcus NagleMSc. Dimitri ArgyropoulosProf. Dr. Joachim Müller

2

Why Lasers?

Low cost and uncomplicated (vs. NIR and hyperspectral imaging)

Continous monitoring possible (In-line)

Avoids product contamination (Non-contact)

Easy handling (More flexible in terms of portability)

Simultaneous prediction of physico-chemical properties

3

Ginger Ginger is the rhizome of the plant Zingiber officinale consumed as a delicacy, medicine and spice. Other notable members of this plant family are turmeric, cardamom and galangal.

FAOSTAT 2010

4

In addition to moisture content and colour changes, oil content will be investigated as well.

In addition to moisture content and colour changes, oil content will be investigated as well.

Properties Fresh Dry

No cell fracture in the tissue studied.

Ruptured cell releases oil and starch grains in surroundingtissue.

Source: M. Noor Azian el al ( 2003)

Oil

Starch

5

Lorentzian Distribution Function

• Analysis of laser backscattering images with image processing software to obtain LD parameters.

• R-value (mean light intensity of each circular band) will be calculated by MLD function .

Peng el al. (2006)

6

Pre-Test:Ginger

Ginger Fresh_535nm

Ginger Fresh_650nmGinger Dry_535nm

MCwb92%

Fre

sh S

liced

(5m

m)

Gin

ger

7

Pre-Test:Ginger MCwb3.4%

Ginger Dry_535nm

Ginger Dry_650nm

Ove

r dr

ying

S

enar

io

8

Failed Tests

MaizeWet_535nm

RiceWet_650nm

9

Objectives Establish laser backscattering image analysis as a valid method to quantify the major quality parameters of ginger during drying.

To find the most sensitive wavelengths to predict physical and chemical properties.

Develop a prototype (In-line laser appratus) for monitoring of product properties in the dryer.

Properties Method

Moisture Content Gravimetric oven method

Colour L*a*b ( Colorimeter)

Ginger Oil Hydrodistillation of essential oils

10

Experimental Procedures• Mono-wavelength laser light

diodes ( 532, 635 & 785 nm)

• Camera and PaxCam software for obtaining scattering images.

• A high-precision convective over-flow dryer with adjustable temperature, humidity and air velocity.

• The dryer is modified by fitting a transparent glass plate over the airstream and product which can be removed for measurement and sampling

11

Experimental Procedure • Electronic balance to measure

continuous mass reduction.

• A laboratory oven is used to calculate the final moisture content.

• Hyrodistillation in Cleavanger appratus.

• Preprocessing tools (peeling/sclicing/washing etc)

• Colorimeter Minolta Chroma-meter

• Softwares for image analysis and statistical evaluation.

12

Preparation and Data Collection Step 1. Pre-drying Treatment

•Washing and peeling

•Slicing* (whole ginger rhizome for oil extraction)

•Pretreatment (optional) Soaking in 0.5% citric acid to prevent enzymatic browning reactions (S.Phoungchandang et al ,2009)

Drying Conditions Program No. Tem. °C RH (%) Air Velocity

(m/s)

59 40°C 29% 0.25 & 0.75

60 60°C 11% 0.25 & 0.75

* Slice size ( TBA)

13

Preparation and Data Collection

Step 2. Obtaining scattering images and measurement of physico-chemical properties

•Moisture content will be calculated (wet basis) at a frequency of 15 minutes until constant mass is reached.

• Oil content every 30 minutes

• Surafce colour every 30 minutes

•Laser scattering images at each wavelength will be obtained corresponding to the timing of the measurement of selected physico-chemical properties

14

Data Analysis and Interpretation

Step 1. Relationship between LD parameters function (R) and MCa.

• Moisture content actual (MCa) of the product will be projected on a scatter diagram against the corresponding R-values.

• A regression function for predicted moisture content (MCp) will be developed with simple regression tools (MCp=f(R))

EXAMPLE

15

Data Analysis and Interpretation Step 2. Correlation analysis (r), coefficient of determination

(r2), and mean squared error (MSE)

After predicting the values of MCp from the calibration model which were developed in the previous step, we can perform a number of statistical analysis and validation techniques to evaluate the accuracy and performance.

Peng el al. (2006)

EXAMPLE

16

Data Analysis and Interpretation

Step 3. Model development and statistical analysis for colour and oil content.

A similar approach as it was performed for moisture content would be followed to develop calibration models (Xp* = f(R)) for predicting colour and ginger oil concentration and then performing the statistical tests to validate the model.

* Xp Predicted product properties

17

Expected Result and Conclusion

• We expect high coefficient of determination (r2 >0.70) for the calibration models.

• The SEP is the standard error of prediction in order to evaluate the effectiveness of each wavelength to predict a certain quality parameter.

• If the model is suitably validated by the different statistical tests and techniques, we can conclude that the laser backscattering image analysis could be an applicable technology in the field of in-line monitoring of drying process for the future.

18

Working Plan Time Period ( April to September 2012)

Tasks and Mile Stones

April Conducting experiments

End of April First set of data analysis

May Conducting experiments

June Conducting experiments

July Data evaluation

August Presenting reuslt and analysisThesis preparation

September Thesis presentation

19

Thank you very much for your kind attention