Shoot-out

Shoot-out (revised 7/14/2008 - see first paragraph)

2008 Winners (posted 8/18/2008)

First Place - Dave Funk, Cathy Brenner, Rich Pierce, USDA-GIPSA
Second Place - Benoit Igne, Iowa State University
Third Place - Dave Ryan, Solae LLC
Fourth Place - Pierre Dardenne, CRA-W

The Software Shoot-out has been a staple of the IDRC. This page provides access to historical sets of shoot-out data, as well as the data set and background information for the 2008 IDRC. The Shoot-out rules published in the first issue of the 2008 edition of the NIR Spectrum have been revised slightly (July 2008) and are replaced with the posting here in pdf file format. (Biggest change - Advance paper requested but no longer required.)

2008 Wheat Functionality as Measured by Diffuse Reflectance
of Whole Grain Canadian Wheat

(contributed by P.C. Williams, PDK Grain; retired from Grain Research Laboratory, Canadian Grain Commission, Winnipeg, Manitoba)

Instrument Data

Samples were scanned on a Foss Model 6500, using a rectangular sample cell, with “clip-on” removable back. The cell was always filled. The instrument was monitored daily using a sealed Check-cell, and a check sample. Both NSAS and WinISI diagnostics were performed most days. Data recorded included wavelength standard, a reference sample, noise test, and bandwidth test. The instrument was never turned off, but the lamp was turned off after use every day, and the instrument allowed to warm up for at least 30 minutes before use.

Spectra anomalies in the spectra occur at about 780-790, and 520-530 nm areas.

Sample Data

Samples are all of whole grain wheat, recorded from Western Canadian Plant-breeders’ material. Reference data were all determined at the Grain Research Laboratory, Winnipeg.

Reference data are Protein content (%); Test weight (Kg/Hl); PSI (wheat kernel texture) (%, w/w); Farinograph water absorption (%, w/w); Farinograph dough development time (minutes); and Farinograph mixing tolerance index [Brabender units (related to torque)]. Most of these are functionality factors, so there are no “classical” absorbers associated with them. For a more detailed description of each property, visit the Canadian Grain Commission's web site (click here; note that the Farinograph properties are listed under Farinogram).

The link following this paragraph directs you to a pdf file that summarizes the calibration set samples (see below for definitions of calibration and validation samples) by record number, sample number, Canadian wheat class, and year. (For more information about the wheat classes, click this here to be taken to the Canadian Grain Commission's descriptions.)

Description and Summarization pdf file

The following two Excel files contain information on the Canadian wheat class for each sample (see the paragraph that follows for an explanation of the two sets). In each spreadsheet file, a column exists for each wheat class. A sample's wheat class is identified by a "1" in its corresponding cell, with all other cells as "0".

ExcelFred2008.xls

ExcelFred99.xls

There are 775 spectra for calibration and 107 for validation. The file Fred2008 should be used for calibration and Fred99 for validation. The samples in Fred2008 represent samples grown in 1998 and 2000-2005. Note that if your version of Fred2008 has 777 spectra, please delete record numbers 680 and 681 (sample numbers 38565 and 38566, respectively), as these are two samples that don't have valid protein content values. Fred99 samples were grown in 1999, and are quite separate from the Fred2008 samples.

The original data were in NSAS format. Dr. Howard Mark converted the data to JCAMP-DX, MATLAB, and Unscrambler formats, then compressed them with WinZip to make them all available for downloading. They are organized according to data format; each format is in a separate zip file, each containing both data sets. The description file is separate.

The zipped files may be downloaded using the following links:

NSAS format - 3.11 MB

JCAMP-DX format - 4.94 MB

MATLAB format - 3.35 MB

Unscrambler format - 2.96 MB

__________________________________________________________________________

Previous Shoot-outs

2006 - Soil properties by diffuse reflectance near-infrared spectroscopy.

2006 page

2004 - Designing the best possible calibration for mid-infrared diffuse reflectance spectra.

2004 page

2002 - Designing the best possible calibration transfer procedure.