Suite 103,56 Kitchner Pde

Bankstown NSW 2200

Tel: 02 9708 5068

Fax: 02 9708 5537

 
NIR Technology Australia

 

Application Note 02: Cropscan 2000G Sorghum Analysis

February 2000

 Introduction
The Cropscan 2000G Whole Grain Analyser has been used to develop a calibration capable of predicting both the protein and moisture content of sorghum samples grown in Australia. Near Infrared (NIR) spectra of sorghum are shown below.

.

Description
52 samples of sorghum were scanned on the Cropscan 2000G between 720-1100nm using a 16mm pathlength cell. Using the software package Unscrambler, a partial least squares (PLS-1) regression was performed on the data to develop a calibration, which relates the spectral data to the prescribed protein and moisture values of the sorghum samples

 

Results
Ten (10) duplicate sorghum samples were used as a prediction set to determine the predictive ability of the calibration model produced. The individual samples were scanned five (5) times and the average was taken as the protein and moisture values. The results are shown below

Sample ID

Protein NIR (1)

Protein Ref (2)

Diff (1) & (2)

Moisture NIR (3)

Moisture Ref (4)

Diff (3) & (4)

1a

10.2

10.1

-0.1

11.4

11.9

0.5

1b

10.2

10.1

-0.1

11.3

11.9

0.6

2a

10.3

10.3

0

12.9

13.2

0.3

2b

10.3

10.3

0

12.9

13.2

0.3

3a

7.9

7.7

-0.2

12.9

12.8

-0.1

3b

8.2

7.7

-0.5

12.8

12.8

0

4a

9.7

9.3

-0.4

13

13.2

0.2

4b

10.1

9.3

-0.8

13

13.2

0.2

5a

9

9.5

0.5

11.9

11.8

-0.1

5b

9

9.5

0.5

11.9

11.8

-0.1

6a

9.3

9.5

0.2

12

12.2

0.2

6b

9.3

9.5

0.2

12.1

12.2

0.1

7a

8.6

8.4

-0.2

12.1

12.1

0

7b

8.6

8.4

-0.2

12.1

12.1

0

8a

9.7

10.3

0.6

11.8

11.9

0.1

8b

9.9

10.3

0.4

11.8

11.9

0.1

9a

9.2

10.1

0.9

13.4

13.4

0

9b

9.3

10.1

0.8

13.5

13.4

-0.1

10a

8

7.5

-0.5

12.4

11.8

-0.6

10b

8

7.5

-0.5

12.5

11.8

-0.7

 

Statistical analyses of the above data are given in the following table

 

Protein Analysis

Moisture Analysis

Bias

0.03

0.04

Standard Deviation

0.47

0.31

SDD*

0.15

0.07

* SDD = Standard Deviation of Differences between duplicate samples

A graphical representation of the data is presented below

 

Conclusion
The Cropscan 2000G can be used to measure the moisture content in sorghum with a good reproducibility (SDD for the duplicate samples of 0.07%) and acceptable accuracy. This provides a method for farmers to confidently deliver sorghum with a moisture content less than 13.5% (which is the maximum allowable level). Also the farmer is potentially able to deliver a more marketable product on the basis of higher protein content (which is can be measured to an accuracy of 0.5%).